diff --git a/added_tokens.json b/added_tokens.json new file mode 100644 index 0000000000000000000000000000000000000000..19727155b7493cab38b90516c4a7eed96ab7412a --- /dev/null +++ b/added_tokens.json @@ -0,0 +1,5 @@ +{ + "<|model|>": 32002, + "<|system|>": 32000, + "<|user|>": 32001 +} diff --git a/config.json b/config.json new file mode 100644 index 0000000000000000000000000000000000000000..37a416a06f53f302aa7a04b9cf7c1acea9afb1fd --- /dev/null +++ b/config.json @@ -0,0 +1,27 @@ +{ + "_name_or_path": "/root/models/tsukasa-120b-adapter/merged", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "bos_token_id": 1, + "eos_token_id": 2, + "hidden_act": "silu", + "hidden_size": 8192, + "initializer_range": 0.02, + "intermediate_size": 28672, + "max_position_embeddings": 4096, + "model_type": "llama", + "num_attention_heads": 64, + "num_hidden_layers": 137, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": null, + "rope_theta": 10000.0, + "tie_word_embeddings": false, + "torch_dtype": "float16", + "transformers_version": "4.35.1", + "use_cache": false, + "vocab_size": 32003 +} diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..ff27dfcd352035fa795d384dac77c8f14f18c8ad --- /dev/null +++ b/generation_config.json @@ -0,0 +1,6 @@ +{ + "_from_model_config": true, + "bos_token_id": 1, + "eos_token_id": 2, + "transformers_version": "4.35.1" +} diff --git a/pytorch_model-00001-of-00050.bin b/pytorch_model-00001-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..a1bc3996b7f0e5abe144a45941cc214cfbe6178c --- /dev/null +++ b/pytorch_model-00001-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:186d6c54ae4279ebf8cff3a5da4231417a46d36ed72a1a6b0e5071c22b7baf3f +size 4718715602 diff --git a/pytorch_model-00002-of-00050.bin b/pytorch_model-00002-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..4d256d734ef12760aa4d4cea1a4d867e57bcbb33 --- /dev/null +++ b/pytorch_model-00002-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab2023748313f417febfce168495128b714dbffd5fc0eb606cca9ed2ad58fedf +size 4664173854 diff --git a/pytorch_model-00003-of-00050.bin b/pytorch_model-00003-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..ad070343a346b3e0fbe4db3f777e7f386443e79c --- /dev/null +++ b/pytorch_model-00003-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5dee0a01470be91507fee66bdaf9c367410b81a1cf250616e02b18033f7228d2 +size 4999718222 diff --git a/pytorch_model-00004-of-00050.bin b/pytorch_model-00004-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..a1445adbc2751dfd157fbacb77b13bec175b6e6d --- /dev/null +++ b/pytorch_model-00004-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:25cbf4d9dc7c23d23c028434dc6922d624bbd87540eb3e09830d01f420a29cee +size 4966163202 diff --git a/pytorch_model-00005-of-00050.bin b/pytorch_model-00005-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..2f49cfb000fba110ee8a07f2ef80c62d33240c67 --- /dev/null +++ b/pytorch_model-00005-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:391d6a7e325ee773af874f1decc1824f242371f42c16645d10933065ced42612 +size 4664140510 diff --git a/pytorch_model-00006-of-00050.bin b/pytorch_model-00006-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..04fda4a119acea42d9e8fad19b0c09244597c20e --- /dev/null +++ b/pytorch_model-00006-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:64faea72ca6df144f907894452b25e6aae8aa169b021d7926df9e146c2bcd5cb +size 4664173918 diff --git a/pytorch_model-00007-of-00050.bin b/pytorch_model-00007-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..c3f3567a207b9e35fb3716b4e3c68d0dba86b842 --- /dev/null +++ b/pytorch_model-00007-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5220808ee358dee63f8c7c7ec037cebc4a00d0721b7f07d787bc83dffe7cbc91 +size 4664173918 diff --git a/pytorch_model-00008-of-00050.bin b/pytorch_model-00008-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..cbc61a0107a2ddf4e05c2ba50ddedf92a6c317b6 --- /dev/null +++ b/pytorch_model-00008-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cfc95644c754e6bb8b658f8af70ca8789585918ca9e80f27fbf8b336923b788d +size 4999718286 diff --git a/pytorch_model-00009-of-00050.bin b/pytorch_model-00009-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..fdae1e6f4efe94cd17e0e337dc4387ece04a435e --- /dev/null +++ b/pytorch_model-00009-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e0c91587fc7d3e47a2aafec6708988ab7a10ed58ca39781f3127cbf097cf6b1 +size 4966163202 diff --git a/pytorch_model-00010-of-00050.bin b/pytorch_model-00010-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..83bda0c0ba5185e908f1625239a17e4916032918 --- /dev/null +++ b/pytorch_model-00010-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6759e1d6771b75b4f3ddf02d30cf37c82de5a863cd6af95c89742e258cf933f +size 4664140510 diff --git a/pytorch_model-00011-of-00050.bin b/pytorch_model-00011-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..4edc3f053cbd362e29eeeb2489e3a18d8f5d4012 --- /dev/null +++ b/pytorch_model-00011-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d1ac0f9bc2b2cbd7dc3c809428e05be1a5a3f55d7c7eeeb36d6664344913fdc9 +size 4664173918 diff --git a/pytorch_model-00012-of-00050.bin b/pytorch_model-00012-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..e04d304119b4f01c43d4bbbbd0be9b1574942d66 --- /dev/null +++ b/pytorch_model-00012-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b23a4b70eb61aa08b31794d852422affae41cdbce6b757cb0744448f04d72740 +size 4664173918 diff --git a/pytorch_model-00013-of-00050.bin b/pytorch_model-00013-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..784722382f522f6e91b050f009402773fcd238bb --- /dev/null +++ b/pytorch_model-00013-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:850632b6cc71044b3ed30401f99e674da646f001a8c0e21baec40dea2005cb78 +size 4999718286 diff --git a/pytorch_model-00014-of-00050.bin b/pytorch_model-00014-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..830458410c37efcd08489b57385c845b2600907f --- /dev/null +++ b/pytorch_model-00014-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7d193fbd9d21d9bf5dec5f6af5273641bad28d2e60ce95c65bee0971f394b3b3 +size 4966163202 diff --git a/pytorch_model-00015-of-00050.bin b/pytorch_model-00015-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..a719f7106a49baa83b3772998b7e89695d83ffe1 --- /dev/null +++ b/pytorch_model-00015-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c2bdc8d8d00eeb39e6edd880c84b6e1b715a2c72e5f0a13137e2414065a1f95 +size 4664140510 diff --git a/pytorch_model-00016-of-00050.bin b/pytorch_model-00016-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..e7a84e46419c72d2efb6e8b17910dfe51cd2d4a7 --- /dev/null +++ b/pytorch_model-00016-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63da89d7812984d3fa612ba42002cf4a74f6c9a8e6b184373cba2bb80bc65ce8 +size 4664173918 diff --git a/pytorch_model-00017-of-00050.bin b/pytorch_model-00017-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..3e8314a40b44c726ebc378dd0b185a2a20498d20 --- /dev/null +++ b/pytorch_model-00017-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7600fa1fb39de29053081a2c922b8a8819751a48f297a57f8527c623234c864e +size 4664173918 diff --git a/pytorch_model-00018-of-00050.bin b/pytorch_model-00018-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..30759a3720c24a114ad0ea533104710b51c41971 --- /dev/null +++ b/pytorch_model-00018-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06f3a4ac7b9b97a159bc9602895412eadfa0e1c11d3e10c82a0859fdbe120381 +size 4999718286 diff --git a/pytorch_model-00019-of-00050.bin b/pytorch_model-00019-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..c154597ef14ce416e31bbebec58c7d06095bd5e7 --- /dev/null +++ b/pytorch_model-00019-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28f3b8409643a46630e601a824f2ccfc254d7e075c8d183725faf7e552b8b5b8 +size 4966163202 diff --git a/pytorch_model-00020-of-00050.bin b/pytorch_model-00020-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..3ed346e9cc13ae60624b8acc716d9ab8f2832624 --- /dev/null +++ b/pytorch_model-00020-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5f7fcb9c0e85a874e841c77c828af86a55c90acb5f1f9dcad85992a642e64668 +size 4664140510 diff --git a/pytorch_model-00021-of-00050.bin b/pytorch_model-00021-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..62e40feca1da38ed5d85b4db201809af6999eaec --- /dev/null +++ b/pytorch_model-00021-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd04279977c48666c8b008c1a1a0c5b7d90fab5a5c768a2c89ba2eb3fdb5e8b3 +size 4664173918 diff --git a/pytorch_model-00022-of-00050.bin b/pytorch_model-00022-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..6f9d09a83c7ca7a56e3c25418a8cfd30695ec70c --- /dev/null +++ b/pytorch_model-00022-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e175ebfa7e3d5de4b6519745b9d466758c7cd45ba797340d48e0446940d2565a +size 4664173918 diff --git a/pytorch_model-00023-of-00050.bin b/pytorch_model-00023-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..2f11296883a01e715256f5d1aed0658ceb760235 --- /dev/null +++ b/pytorch_model-00023-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c3972f9f85e143190bfdfbead2e46983fb5be4ebd2437cca8ec125f30c56fc90 +size 4999718286 diff --git a/pytorch_model-00024-of-00050.bin b/pytorch_model-00024-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..4649d27ddbbafb429f69985209188f5b4f7ac2d9 --- /dev/null +++ b/pytorch_model-00024-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1483c46205452d5697ba7da21248b17486a51b960ca7369f6b675f4a63da1183 +size 4966163202 diff --git a/pytorch_model-00025-of-00050.bin b/pytorch_model-00025-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..7bcd8fc7e88404df01df9d570e254b9019eef9b3 --- /dev/null +++ b/pytorch_model-00025-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3da0d86c1940a68afe01663ac5ac30c806cf94e0a6cf305b187cc5aa77a9a69b +size 4664140510 diff --git a/pytorch_model-00026-of-00050.bin b/pytorch_model-00026-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..66e055489b0c4a3ec0d337639ff1192a8b120751 --- /dev/null +++ b/pytorch_model-00026-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6021780e9be4b68ac08905abe0ce993e082857a03e3966b1ae7667fa120c1d27 +size 4664173918 diff --git a/pytorch_model-00027-of-00050.bin b/pytorch_model-00027-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..a0c5368b6b2b254d6fba3a96476a51c01efb1ad6 --- /dev/null +++ b/pytorch_model-00027-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c1584606149c5d95014c4d2868eb11ad8b5df21ea0dafbe2e75d4148be0c991 +size 4664173918 diff --git a/pytorch_model-00028-of-00050.bin b/pytorch_model-00028-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..b81467bc878c5fdca1911265c3c32d4d3a7c5e95 --- /dev/null +++ b/pytorch_model-00028-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad20527210677734e2bd977a3b73a3602788151eca6ae93027dc9dbadf373dbc +size 4999718286 diff --git a/pytorch_model-00029-of-00050.bin b/pytorch_model-00029-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..215716aa30214922c69603ba6404b5e255b936d4 --- /dev/null +++ b/pytorch_model-00029-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7df906fb6de5f714c431e4539a176b45d22b7a5e92193d2a47bd94f5e6eacf58 +size 4966163202 diff --git a/pytorch_model-00030-of-00050.bin b/pytorch_model-00030-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..26f3921b7abad2dc2cd20ae31be4a5903d61687e --- /dev/null +++ b/pytorch_model-00030-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ffd0039b8a7bcac365327335113cc8fac7987b8307f8491556a4151fb07b6e3a +size 4664140510 diff --git a/pytorch_model-00031-of-00050.bin b/pytorch_model-00031-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..759328f1fabb21e7f0618b588f9373b75fb70ad6 --- /dev/null +++ b/pytorch_model-00031-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb907e077c13e6151d3d476e0941318d20352c9a0bd6c99fa0f87d66b5f3d640 +size 4664173918 diff --git a/pytorch_model-00032-of-00050.bin b/pytorch_model-00032-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..8dbcdd0d5744b075219b1d4c4073c97f8a060219 --- /dev/null +++ b/pytorch_model-00032-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b2e8f6e73b0d96997c574c77bf8280d8215f75c8122bdd041d002ca937c8665 +size 4664173918 diff --git a/pytorch_model-00033-of-00050.bin b/pytorch_model-00033-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..b3fed29c7d1e10e1a14b127149e17692b795af33 --- /dev/null +++ b/pytorch_model-00033-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9026feedef618e74351dbbb68963d806a9bb981ffd97491afcfe2c12a5d5b989 +size 4999718286 diff --git a/pytorch_model-00034-of-00050.bin b/pytorch_model-00034-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..2c4ca1bf029895bf4c2bccbbae712d76685c07dc --- /dev/null +++ b/pytorch_model-00034-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e503fece969bd49ded7f92b6e48e5d90b4420865e82b61aa7317c7a5dc53c1f +size 4966163202 diff --git a/pytorch_model-00035-of-00050.bin b/pytorch_model-00035-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..a3905eef592f28619ae875a9aec0646696f5d149 --- /dev/null +++ b/pytorch_model-00035-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e75227aa3905d3ac8607be693b60707122d1480bb9f92046518d5ae99a1ab9e +size 4664140510 diff --git a/pytorch_model-00036-of-00050.bin b/pytorch_model-00036-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..60a90e29c5c58ed863eb021d3ae92ec9fc750cd2 --- /dev/null +++ b/pytorch_model-00036-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:77dbad3edc54a735122a2e6dcb16d6691ea3538e90990d2d26dd6b352af41051 +size 4664173918 diff --git a/pytorch_model-00037-of-00050.bin b/pytorch_model-00037-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..5fef8ccb1f9b5f2907bb815f8afc853c90bf63ba --- /dev/null +++ b/pytorch_model-00037-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5b2d30433bbd8d7a771c30a59a39ae30ae27732e2977cc1c39fa059336ffec71 +size 4664173918 diff --git a/pytorch_model-00038-of-00050.bin b/pytorch_model-00038-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..5cd1b0a45e2e03b5e9c109dc0a58ab90ee8a5172 --- /dev/null +++ b/pytorch_model-00038-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d96fa16d395893bdb25d2e13d45b8a946acf4dbd91c9f751b3b268dcc7627b29 +size 4999718286 diff --git a/pytorch_model-00039-of-00050.bin b/pytorch_model-00039-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..31d591fbef01fd138cad6ab72534ce755b6971c5 --- /dev/null +++ b/pytorch_model-00039-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1f1bbf210dcb66f622d4134128dbc93319b1cadd3f41e7888f4709d57c879444 +size 4966163266 diff --git a/pytorch_model-00040-of-00050.bin b/pytorch_model-00040-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..c93ba03b77008839a544c0b83519a9c3e1396b81 --- /dev/null +++ b/pytorch_model-00040-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3cf46ede853326434b706949c4261f801d6fa134aad0cb8405b771c6e049b52 +size 4664140574 diff --git a/pytorch_model-00041-of-00050.bin b/pytorch_model-00041-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..4f1e3e036fb12048b331ab07fdfc983c3b2f0226 --- /dev/null +++ b/pytorch_model-00041-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1f1426f9c333566c359c8cf6e2d32a7bd5d8b94222c28fd5b2a3d23d4051855c +size 4664173918 diff --git a/pytorch_model-00042-of-00050.bin b/pytorch_model-00042-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..4050e4e02bb60ef2ba94b45520a3873ce49692d1 --- /dev/null +++ b/pytorch_model-00042-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f768f6958d0e73cf965cfcde9f83f6c0079d23c8be71699cae7db5054d2bce84 +size 4664173918 diff --git a/pytorch_model-00043-of-00050.bin b/pytorch_model-00043-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..39b27f1ae7fd871b44cd1ccf3585b4e962889116 --- /dev/null +++ b/pytorch_model-00043-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dcd372000479f8c501ed1fe1c9078db4b37bb848d0021c24449ddf093b66232e +size 4999718286 diff --git a/pytorch_model-00044-of-00050.bin b/pytorch_model-00044-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..67e8b25865910faabdd9354c3de5f360b9347050 --- /dev/null +++ b/pytorch_model-00044-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9fe5c6d1d8ba820eca02b152c791d3cbd80ee5aca35120d5f220bef62692120c +size 4966163266 diff --git a/pytorch_model-00045-of-00050.bin b/pytorch_model-00045-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..fc00d5f84198df2ceb7118a8cd08e1a4135e816a --- /dev/null +++ b/pytorch_model-00045-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:503dc7636f64a38756c93e73b5f4d96bb8aea28e0d38c9a61b9a43c9bac2a6e4 +size 4664140574 diff --git a/pytorch_model-00046-of-00050.bin b/pytorch_model-00046-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..ddd291146a8d5772df5c434167f23df411ad32cc --- /dev/null +++ b/pytorch_model-00046-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:46f40ffd72d8646f849b59dd9949a199fc72db36ab457ed3e9aab2b0ab7ba1a8 +size 4664173918 diff --git a/pytorch_model-00047-of-00050.bin b/pytorch_model-00047-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..7bf43eafbab87e1b571f9848773b336b06d81ea6 --- /dev/null +++ b/pytorch_model-00047-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0c1c3f7152fd607ecf783fcaccb34bd80f8c85669adcd67e074c232b71a4b50 +size 4664173918 diff --git a/pytorch_model-00048-of-00050.bin b/pytorch_model-00048-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..49aab00e603c5b66e91d30ca723ed791f8bcd6dc --- /dev/null +++ b/pytorch_model-00048-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe3ff5475a8d54979087a993b69d589f6d83a68af1cf067ac5cb08d0802e5212 +size 4999718286 diff --git a/pytorch_model-00049-of-00050.bin b/pytorch_model-00049-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..b02300ab6fa18624e726fca1bec1f5b080200538 --- /dev/null +++ b/pytorch_model-00049-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e8a1a04199e0531a1b3dd9507f1056b2822c65648d91fb888b514fcfdb8daee +size 4966179924 diff --git a/pytorch_model-00050-of-00050.bin b/pytorch_model-00050-of-00050.bin new file mode 100644 index 0000000000000000000000000000000000000000..19f27d2b766e5f57e357b79293d306f2e4dbeebf --- /dev/null +++ b/pytorch_model-00050-of-00050.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6e2648958db8a29cf50d92b62e7dd4ff880ba4b4292d0cad9ab50dccc6147d88 +size 524338565 diff --git a/pytorch_model.bin.index.json b/pytorch_model.bin.index.json new file mode 100644 index 0000000000000000000000000000000000000000..9908f2b1b058a9c41a7a36d0ede10fb61bcfc780 --- /dev/null +++ b/pytorch_model.bin.index.json @@ -0,0 +1,1243 @@ +{ + "metadata": { + "total_size": 235497996288 + }, + "weight_map": { + "lm_head.weight": "pytorch_model-00050-of-00050.bin", + "model.embed_tokens.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.10.input_layernorm.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.mlp.down_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.mlp.up_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.100.input_layernorm.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.100.mlp.down_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.100.mlp.gate_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.100.mlp.up_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.100.post_attention_layernorm.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.100.self_attn.k_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.100.self_attn.o_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.100.self_attn.q_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.100.self_attn.v_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.101.input_layernorm.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.mlp.down_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.mlp.gate_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.mlp.up_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.post_attention_layernorm.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.self_attn.k_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.self_attn.o_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.self_attn.q_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.101.self_attn.v_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.input_layernorm.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.mlp.down_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.mlp.gate_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.mlp.up_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.post_attention_layernorm.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.self_attn.k_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.self_attn.o_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.self_attn.q_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.102.self_attn.v_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.103.input_layernorm.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.103.mlp.down_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.103.mlp.gate_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.103.mlp.up_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.103.post_attention_layernorm.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.103.self_attn.k_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.103.self_attn.o_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.103.self_attn.q_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.103.self_attn.v_proj.weight": "pytorch_model-00037-of-00050.bin", + "model.layers.104.input_layernorm.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.mlp.down_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.mlp.gate_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.mlp.up_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.post_attention_layernorm.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.self_attn.k_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.self_attn.o_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.self_attn.q_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.104.self_attn.v_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.input_layernorm.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.mlp.down_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.mlp.gate_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.mlp.up_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.post_attention_layernorm.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.self_attn.k_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.self_attn.o_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.self_attn.q_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.105.self_attn.v_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.106.input_layernorm.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.106.mlp.down_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.106.mlp.gate_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.106.mlp.up_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.106.post_attention_layernorm.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.106.self_attn.k_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.106.self_attn.o_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.106.self_attn.q_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.106.self_attn.v_proj.weight": "pytorch_model-00038-of-00050.bin", + "model.layers.107.input_layernorm.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.mlp.down_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.mlp.gate_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.mlp.up_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.post_attention_layernorm.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.self_attn.k_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.self_attn.o_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.self_attn.q_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.107.self_attn.v_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.input_layernorm.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.mlp.down_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.mlp.gate_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.mlp.up_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.post_attention_layernorm.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.self_attn.k_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.self_attn.o_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.self_attn.q_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.108.self_attn.v_proj.weight": "pytorch_model-00039-of-00050.bin", + "model.layers.109.input_layernorm.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.mlp.down_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.mlp.gate_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.mlp.up_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.post_attention_layernorm.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.self_attn.k_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.self_attn.o_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.self_attn.q_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.109.self_attn.v_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.11.input_layernorm.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.mlp.down_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.mlp.up_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.110.input_layernorm.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.mlp.down_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.mlp.gate_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.mlp.up_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.post_attention_layernorm.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.self_attn.k_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.self_attn.o_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.self_attn.q_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.110.self_attn.v_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.111.input_layernorm.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.111.mlp.down_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.111.mlp.gate_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.111.mlp.up_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.111.post_attention_layernorm.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.111.self_attn.k_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.111.self_attn.o_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.111.self_attn.q_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.111.self_attn.v_proj.weight": "pytorch_model-00040-of-00050.bin", + "model.layers.112.input_layernorm.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.mlp.down_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.mlp.gate_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.mlp.up_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.post_attention_layernorm.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.self_attn.k_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.self_attn.o_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.self_attn.q_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.112.self_attn.v_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.input_layernorm.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.mlp.down_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.mlp.gate_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.mlp.up_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.post_attention_layernorm.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.self_attn.k_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.self_attn.o_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.self_attn.q_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.113.self_attn.v_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.114.input_layernorm.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.114.mlp.down_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.114.mlp.gate_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.114.mlp.up_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.114.post_attention_layernorm.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.114.self_attn.k_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.114.self_attn.o_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.114.self_attn.q_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.114.self_attn.v_proj.weight": "pytorch_model-00041-of-00050.bin", + "model.layers.115.input_layernorm.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.mlp.down_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.mlp.gate_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.mlp.up_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.post_attention_layernorm.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.self_attn.k_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.self_attn.o_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.self_attn.q_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.115.self_attn.v_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.input_layernorm.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.mlp.down_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.mlp.gate_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.mlp.up_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.post_attention_layernorm.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.self_attn.k_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.self_attn.o_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.self_attn.q_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.116.self_attn.v_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.117.input_layernorm.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.117.mlp.down_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.117.mlp.gate_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.117.mlp.up_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.117.post_attention_layernorm.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.117.self_attn.k_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.117.self_attn.o_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.117.self_attn.q_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.117.self_attn.v_proj.weight": "pytorch_model-00042-of-00050.bin", + "model.layers.118.input_layernorm.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.mlp.down_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.mlp.gate_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.mlp.up_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.post_attention_layernorm.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.self_attn.k_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.self_attn.o_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.self_attn.q_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.118.self_attn.v_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.input_layernorm.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.mlp.down_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.mlp.gate_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.mlp.up_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.post_attention_layernorm.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.self_attn.k_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.self_attn.o_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.self_attn.q_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.119.self_attn.v_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.12.input_layernorm.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.mlp.down_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.mlp.up_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.120.input_layernorm.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.120.mlp.down_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.120.mlp.gate_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.120.mlp.up_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.120.post_attention_layernorm.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.120.self_attn.k_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.120.self_attn.o_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.120.self_attn.q_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.120.self_attn.v_proj.weight": "pytorch_model-00043-of-00050.bin", + "model.layers.121.input_layernorm.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.mlp.down_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.mlp.gate_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.mlp.up_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.post_attention_layernorm.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.self_attn.k_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.self_attn.o_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.self_attn.q_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.121.self_attn.v_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.input_layernorm.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.mlp.down_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.mlp.gate_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.mlp.up_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.post_attention_layernorm.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.self_attn.k_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.self_attn.o_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.self_attn.q_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.122.self_attn.v_proj.weight": "pytorch_model-00044-of-00050.bin", + "model.layers.123.input_layernorm.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.mlp.down_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.mlp.gate_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.mlp.up_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.post_attention_layernorm.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.self_attn.k_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.self_attn.o_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.self_attn.q_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.123.self_attn.v_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.input_layernorm.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.mlp.down_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.mlp.gate_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.mlp.up_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.post_attention_layernorm.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.self_attn.k_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.self_attn.o_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.self_attn.q_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.124.self_attn.v_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.125.input_layernorm.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.125.mlp.down_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.125.mlp.gate_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.125.mlp.up_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.125.post_attention_layernorm.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.125.self_attn.k_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.125.self_attn.o_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.125.self_attn.q_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.125.self_attn.v_proj.weight": "pytorch_model-00045-of-00050.bin", + "model.layers.126.input_layernorm.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.mlp.down_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.mlp.gate_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.mlp.up_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.post_attention_layernorm.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.self_attn.k_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.self_attn.o_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.self_attn.q_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.126.self_attn.v_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.input_layernorm.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.mlp.down_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.mlp.gate_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.mlp.up_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.post_attention_layernorm.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.self_attn.k_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.self_attn.o_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.self_attn.q_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.127.self_attn.v_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.128.input_layernorm.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.128.mlp.down_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.128.mlp.gate_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.128.mlp.up_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.128.post_attention_layernorm.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.128.self_attn.k_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.128.self_attn.o_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.128.self_attn.q_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.128.self_attn.v_proj.weight": "pytorch_model-00046-of-00050.bin", + "model.layers.129.input_layernorm.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.mlp.down_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.mlp.gate_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.mlp.up_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.post_attention_layernorm.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.self_attn.k_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.self_attn.o_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.self_attn.q_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.129.self_attn.v_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.13.input_layernorm.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.13.mlp.down_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.13.mlp.up_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00005-of-00050.bin", + "model.layers.130.input_layernorm.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.mlp.down_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.mlp.gate_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.mlp.up_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.post_attention_layernorm.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.self_attn.k_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.self_attn.o_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.self_attn.q_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.130.self_attn.v_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.131.input_layernorm.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.131.mlp.down_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.131.mlp.gate_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.131.mlp.up_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.131.post_attention_layernorm.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.131.self_attn.k_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.131.self_attn.o_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.131.self_attn.q_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.131.self_attn.v_proj.weight": "pytorch_model-00047-of-00050.bin", + "model.layers.132.input_layernorm.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.mlp.down_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.mlp.gate_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.mlp.up_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.post_attention_layernorm.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.self_attn.k_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.self_attn.o_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.self_attn.q_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.132.self_attn.v_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.input_layernorm.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.mlp.down_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.mlp.gate_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.mlp.up_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.post_attention_layernorm.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.self_attn.k_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.self_attn.o_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.self_attn.q_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.133.self_attn.v_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.134.input_layernorm.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.134.mlp.down_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.134.mlp.gate_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.134.mlp.up_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.134.post_attention_layernorm.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.134.self_attn.k_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.134.self_attn.o_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.134.self_attn.q_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.134.self_attn.v_proj.weight": "pytorch_model-00048-of-00050.bin", + "model.layers.135.input_layernorm.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.mlp.down_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.mlp.gate_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.mlp.up_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.post_attention_layernorm.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.self_attn.k_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.self_attn.o_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.self_attn.q_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.135.self_attn.v_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.input_layernorm.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.mlp.down_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.mlp.gate_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.mlp.up_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.post_attention_layernorm.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.self_attn.k_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.self_attn.o_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.self_attn.q_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.136.self_attn.v_proj.weight": "pytorch_model-00049-of-00050.bin", + "model.layers.14.input_layernorm.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.mlp.down_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.mlp.up_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.input_layernorm.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.mlp.down_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.mlp.up_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.16.input_layernorm.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.16.mlp.down_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.16.mlp.up_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00006-of-00050.bin", + "model.layers.17.input_layernorm.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.mlp.down_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.mlp.up_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.input_layernorm.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.mlp.down_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.mlp.up_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.19.input_layernorm.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.19.mlp.down_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.19.mlp.up_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00007-of-00050.bin", + "model.layers.2.input_layernorm.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.2.mlp.down_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.2.mlp.up_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00050.bin", + "model.layers.20.input_layernorm.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.mlp.down_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.mlp.up_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.input_layernorm.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.mlp.down_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.mlp.up_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.22.input_layernorm.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.22.mlp.down_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.22.mlp.up_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00008-of-00050.bin", + "model.layers.23.input_layernorm.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.mlp.down_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.mlp.up_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.input_layernorm.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.mlp.down_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.mlp.up_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00009-of-00050.bin", + "model.layers.25.input_layernorm.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.mlp.down_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.mlp.up_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.input_layernorm.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.mlp.down_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.mlp.up_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.27.input_layernorm.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.27.mlp.down_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.27.mlp.up_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00010-of-00050.bin", + "model.layers.28.input_layernorm.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.mlp.down_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.mlp.up_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.input_layernorm.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.mlp.down_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.mlp.up_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.3.input_layernorm.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.mlp.down_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.mlp.up_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.30.input_layernorm.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.30.mlp.down_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.30.mlp.up_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00011-of-00050.bin", + "model.layers.31.input_layernorm.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.mlp.down_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.mlp.up_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.input_layernorm.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.mlp.down_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.mlp.up_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.self_attn.k_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.self_attn.q_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.32.self_attn.v_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.33.input_layernorm.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.33.mlp.down_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.33.mlp.up_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.33.self_attn.k_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.33.self_attn.q_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.33.self_attn.v_proj.weight": "pytorch_model-00012-of-00050.bin", + "model.layers.34.input_layernorm.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.mlp.down_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.mlp.up_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.self_attn.k_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.self_attn.q_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.34.self_attn.v_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.input_layernorm.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.mlp.down_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.mlp.up_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.self_attn.k_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.self_attn.q_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.35.self_attn.v_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.36.input_layernorm.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.36.mlp.down_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.36.mlp.up_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.36.self_attn.k_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.36.self_attn.q_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.36.self_attn.v_proj.weight": "pytorch_model-00013-of-00050.bin", + "model.layers.37.input_layernorm.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.mlp.down_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.mlp.up_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.self_attn.k_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.self_attn.q_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.37.self_attn.v_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.input_layernorm.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.mlp.down_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.mlp.up_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.self_attn.k_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.self_attn.q_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.38.self_attn.v_proj.weight": "pytorch_model-00014-of-00050.bin", + "model.layers.39.input_layernorm.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.mlp.down_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.mlp.up_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.self_attn.k_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.self_attn.q_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.39.self_attn.v_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.4.input_layernorm.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.mlp.down_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.mlp.up_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.40.input_layernorm.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.mlp.down_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.mlp.gate_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.mlp.up_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.post_attention_layernorm.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.self_attn.k_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.self_attn.o_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.self_attn.q_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.40.self_attn.v_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.41.input_layernorm.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.41.mlp.down_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.41.mlp.gate_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.41.mlp.up_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.41.post_attention_layernorm.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.41.self_attn.k_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.41.self_attn.o_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.41.self_attn.q_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.41.self_attn.v_proj.weight": "pytorch_model-00015-of-00050.bin", + "model.layers.42.input_layernorm.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.mlp.down_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.mlp.gate_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.mlp.up_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.post_attention_layernorm.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.self_attn.k_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.self_attn.o_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.self_attn.q_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.42.self_attn.v_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.input_layernorm.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.mlp.down_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.mlp.gate_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.mlp.up_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.post_attention_layernorm.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.self_attn.k_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.self_attn.o_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.self_attn.q_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.43.self_attn.v_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.44.input_layernorm.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.44.mlp.down_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.44.mlp.gate_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.44.mlp.up_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.44.post_attention_layernorm.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.44.self_attn.k_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.44.self_attn.o_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.44.self_attn.q_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.44.self_attn.v_proj.weight": "pytorch_model-00016-of-00050.bin", + "model.layers.45.input_layernorm.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.mlp.down_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.mlp.gate_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.mlp.up_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.post_attention_layernorm.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.self_attn.k_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.self_attn.o_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.self_attn.q_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.45.self_attn.v_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.input_layernorm.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.mlp.down_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.mlp.gate_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.mlp.up_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.post_attention_layernorm.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.self_attn.k_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.self_attn.o_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.self_attn.q_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.46.self_attn.v_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.47.input_layernorm.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.47.mlp.down_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.47.mlp.gate_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.47.mlp.up_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.47.post_attention_layernorm.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.47.self_attn.k_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.47.self_attn.o_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.47.self_attn.q_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.47.self_attn.v_proj.weight": "pytorch_model-00017-of-00050.bin", + "model.layers.48.input_layernorm.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.mlp.down_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.mlp.gate_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.mlp.up_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.post_attention_layernorm.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.self_attn.k_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.self_attn.o_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.self_attn.q_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.48.self_attn.v_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.input_layernorm.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.mlp.down_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.mlp.gate_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.mlp.up_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.post_attention_layernorm.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.self_attn.k_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.self_attn.o_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.self_attn.q_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.49.self_attn.v_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.5.input_layernorm.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.5.mlp.down_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.5.mlp.up_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00002-of-00050.bin", + "model.layers.50.input_layernorm.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.50.mlp.down_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.50.mlp.gate_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.50.mlp.up_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.50.post_attention_layernorm.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.50.self_attn.k_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.50.self_attn.o_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.50.self_attn.q_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.50.self_attn.v_proj.weight": "pytorch_model-00018-of-00050.bin", + "model.layers.51.input_layernorm.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.mlp.down_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.mlp.gate_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.mlp.up_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.post_attention_layernorm.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.self_attn.k_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.self_attn.o_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.self_attn.q_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.51.self_attn.v_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.input_layernorm.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.mlp.down_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.mlp.gate_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.mlp.up_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.post_attention_layernorm.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.self_attn.k_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.self_attn.o_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.self_attn.q_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.52.self_attn.v_proj.weight": "pytorch_model-00019-of-00050.bin", + "model.layers.53.input_layernorm.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.mlp.down_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.mlp.gate_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.mlp.up_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.post_attention_layernorm.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.self_attn.k_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.self_attn.o_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.self_attn.q_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.53.self_attn.v_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.input_layernorm.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.mlp.down_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.mlp.gate_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.mlp.up_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.post_attention_layernorm.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.self_attn.k_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.self_attn.o_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.self_attn.q_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.54.self_attn.v_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.55.input_layernorm.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.55.mlp.down_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.55.mlp.gate_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.55.mlp.up_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.55.post_attention_layernorm.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.55.self_attn.k_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.55.self_attn.o_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.55.self_attn.q_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.55.self_attn.v_proj.weight": "pytorch_model-00020-of-00050.bin", + "model.layers.56.input_layernorm.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.mlp.down_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.mlp.gate_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.mlp.up_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.post_attention_layernorm.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.self_attn.k_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.self_attn.o_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.self_attn.q_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.56.self_attn.v_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.input_layernorm.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.mlp.down_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.mlp.gate_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.mlp.up_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.post_attention_layernorm.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.self_attn.k_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.self_attn.o_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.self_attn.q_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.57.self_attn.v_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.58.input_layernorm.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.58.mlp.down_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.58.mlp.gate_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.58.mlp.up_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.58.post_attention_layernorm.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.58.self_attn.k_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.58.self_attn.o_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.58.self_attn.q_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.58.self_attn.v_proj.weight": "pytorch_model-00021-of-00050.bin", + "model.layers.59.input_layernorm.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.mlp.down_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.mlp.gate_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.mlp.up_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.post_attention_layernorm.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.self_attn.k_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.self_attn.o_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.self_attn.q_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.59.self_attn.v_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.6.input_layernorm.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.mlp.down_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.mlp.up_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.60.input_layernorm.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.mlp.down_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.mlp.gate_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.mlp.up_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.post_attention_layernorm.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.self_attn.k_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.self_attn.o_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.self_attn.q_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.60.self_attn.v_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.61.input_layernorm.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.61.mlp.down_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.61.mlp.gate_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.61.mlp.up_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.61.post_attention_layernorm.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.61.self_attn.k_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.61.self_attn.o_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.61.self_attn.q_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.61.self_attn.v_proj.weight": "pytorch_model-00022-of-00050.bin", + "model.layers.62.input_layernorm.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.mlp.down_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.mlp.gate_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.mlp.up_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.post_attention_layernorm.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.self_attn.k_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.self_attn.o_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.self_attn.q_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.62.self_attn.v_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.input_layernorm.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.mlp.down_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.mlp.gate_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.mlp.up_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.post_attention_layernorm.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.self_attn.k_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.self_attn.o_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.self_attn.q_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.63.self_attn.v_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.64.input_layernorm.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.64.mlp.down_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.64.mlp.gate_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.64.mlp.up_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.64.post_attention_layernorm.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.64.self_attn.k_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.64.self_attn.o_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.64.self_attn.q_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.64.self_attn.v_proj.weight": "pytorch_model-00023-of-00050.bin", + "model.layers.65.input_layernorm.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.mlp.down_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.mlp.gate_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.mlp.up_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.post_attention_layernorm.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.self_attn.k_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.self_attn.o_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.self_attn.q_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.65.self_attn.v_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.input_layernorm.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.mlp.down_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.mlp.gate_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.mlp.up_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.post_attention_layernorm.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.self_attn.k_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.self_attn.o_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.self_attn.q_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.66.self_attn.v_proj.weight": "pytorch_model-00024-of-00050.bin", + "model.layers.67.input_layernorm.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.mlp.down_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.mlp.gate_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.mlp.up_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.post_attention_layernorm.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.self_attn.k_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.self_attn.o_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.self_attn.q_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.67.self_attn.v_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.input_layernorm.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.mlp.down_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.mlp.gate_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.mlp.up_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.post_attention_layernorm.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.self_attn.k_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.self_attn.o_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.self_attn.q_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.68.self_attn.v_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.69.input_layernorm.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.69.mlp.down_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.69.mlp.gate_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.69.mlp.up_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.69.post_attention_layernorm.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.69.self_attn.k_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.69.self_attn.o_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.69.self_attn.q_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.69.self_attn.v_proj.weight": "pytorch_model-00025-of-00050.bin", + "model.layers.7.input_layernorm.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.mlp.down_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.mlp.up_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.70.input_layernorm.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.mlp.down_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.mlp.gate_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.mlp.up_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.post_attention_layernorm.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.self_attn.k_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.self_attn.o_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.self_attn.q_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.70.self_attn.v_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.input_layernorm.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.mlp.down_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.mlp.gate_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.mlp.up_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.post_attention_layernorm.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.self_attn.k_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.self_attn.o_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.self_attn.q_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.71.self_attn.v_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.72.input_layernorm.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.72.mlp.down_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.72.mlp.gate_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.72.mlp.up_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.72.post_attention_layernorm.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.72.self_attn.k_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.72.self_attn.o_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.72.self_attn.q_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.72.self_attn.v_proj.weight": "pytorch_model-00026-of-00050.bin", + "model.layers.73.input_layernorm.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.mlp.down_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.mlp.gate_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.mlp.up_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.post_attention_layernorm.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.self_attn.k_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.self_attn.o_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.self_attn.q_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.73.self_attn.v_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.input_layernorm.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.mlp.down_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.mlp.gate_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.mlp.up_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.post_attention_layernorm.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.self_attn.k_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.self_attn.o_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.self_attn.q_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.74.self_attn.v_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.75.input_layernorm.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.75.mlp.down_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.75.mlp.gate_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.75.mlp.up_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.75.post_attention_layernorm.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.75.self_attn.k_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.75.self_attn.o_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.75.self_attn.q_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.75.self_attn.v_proj.weight": "pytorch_model-00027-of-00050.bin", + "model.layers.76.input_layernorm.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.mlp.down_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.mlp.gate_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.mlp.up_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.post_attention_layernorm.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.self_attn.k_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.self_attn.o_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.self_attn.q_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.76.self_attn.v_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.input_layernorm.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.mlp.down_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.mlp.gate_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.mlp.up_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.post_attention_layernorm.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.self_attn.k_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.self_attn.o_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.self_attn.q_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.77.self_attn.v_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.78.input_layernorm.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.78.mlp.down_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.78.mlp.gate_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.78.mlp.up_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.78.post_attention_layernorm.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.78.self_attn.k_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.78.self_attn.o_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.78.self_attn.q_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.78.self_attn.v_proj.weight": "pytorch_model-00028-of-00050.bin", + "model.layers.79.input_layernorm.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.mlp.down_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.mlp.gate_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.mlp.up_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.post_attention_layernorm.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.self_attn.k_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.self_attn.o_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.self_attn.q_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.79.self_attn.v_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.8.input_layernorm.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.8.mlp.down_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.8.mlp.up_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00003-of-00050.bin", + "model.layers.80.input_layernorm.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.mlp.down_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.mlp.gate_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.mlp.up_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.post_attention_layernorm.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.self_attn.k_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.self_attn.o_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.self_attn.q_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.80.self_attn.v_proj.weight": "pytorch_model-00029-of-00050.bin", + "model.layers.81.input_layernorm.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.mlp.down_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.mlp.gate_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.mlp.up_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.post_attention_layernorm.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.self_attn.k_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.self_attn.o_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.self_attn.q_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.81.self_attn.v_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.input_layernorm.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.mlp.down_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.mlp.gate_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.mlp.up_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.post_attention_layernorm.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.self_attn.k_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.self_attn.o_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.self_attn.q_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.82.self_attn.v_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.83.input_layernorm.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.83.mlp.down_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.83.mlp.gate_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.83.mlp.up_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.83.post_attention_layernorm.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.83.self_attn.k_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.83.self_attn.o_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.83.self_attn.q_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.83.self_attn.v_proj.weight": "pytorch_model-00030-of-00050.bin", + "model.layers.84.input_layernorm.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.mlp.down_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.mlp.gate_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.mlp.up_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.post_attention_layernorm.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.self_attn.k_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.self_attn.o_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.self_attn.q_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.84.self_attn.v_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.input_layernorm.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.mlp.down_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.mlp.gate_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.mlp.up_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.post_attention_layernorm.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.self_attn.k_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.self_attn.o_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.self_attn.q_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.85.self_attn.v_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.86.input_layernorm.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.86.mlp.down_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.86.mlp.gate_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.86.mlp.up_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.86.post_attention_layernorm.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.86.self_attn.k_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.86.self_attn.o_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.86.self_attn.q_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.86.self_attn.v_proj.weight": "pytorch_model-00031-of-00050.bin", + "model.layers.87.input_layernorm.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.mlp.down_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.mlp.gate_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.mlp.up_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.post_attention_layernorm.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.self_attn.k_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.self_attn.o_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.self_attn.q_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.87.self_attn.v_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.input_layernorm.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.mlp.down_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.mlp.gate_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.mlp.up_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.post_attention_layernorm.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.self_attn.k_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.self_attn.o_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.self_attn.q_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.88.self_attn.v_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.89.input_layernorm.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.89.mlp.down_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.89.mlp.gate_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.89.mlp.up_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.89.post_attention_layernorm.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.89.self_attn.k_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.89.self_attn.o_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.89.self_attn.q_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.89.self_attn.v_proj.weight": "pytorch_model-00032-of-00050.bin", + "model.layers.9.input_layernorm.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.mlp.down_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.mlp.up_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00004-of-00050.bin", + "model.layers.90.input_layernorm.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.mlp.down_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.mlp.gate_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.mlp.up_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.post_attention_layernorm.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.self_attn.k_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.self_attn.o_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.self_attn.q_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.90.self_attn.v_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.input_layernorm.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.mlp.down_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.mlp.gate_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.mlp.up_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.post_attention_layernorm.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.self_attn.k_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.self_attn.o_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.self_attn.q_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.91.self_attn.v_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.92.input_layernorm.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.92.mlp.down_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.92.mlp.gate_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.92.mlp.up_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.92.post_attention_layernorm.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.92.self_attn.k_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.92.self_attn.o_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.92.self_attn.q_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.92.self_attn.v_proj.weight": "pytorch_model-00033-of-00050.bin", + "model.layers.93.input_layernorm.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.mlp.down_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.mlp.gate_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.mlp.up_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.post_attention_layernorm.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.self_attn.k_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.self_attn.o_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.self_attn.q_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.93.self_attn.v_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.input_layernorm.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.mlp.down_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.mlp.gate_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.mlp.up_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.post_attention_layernorm.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.self_attn.k_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.self_attn.o_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.self_attn.q_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.94.self_attn.v_proj.weight": "pytorch_model-00034-of-00050.bin", + "model.layers.95.input_layernorm.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.mlp.down_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.mlp.gate_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.mlp.up_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.post_attention_layernorm.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.self_attn.k_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.self_attn.o_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.self_attn.q_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.95.self_attn.v_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.input_layernorm.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.mlp.down_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.mlp.gate_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.mlp.up_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.post_attention_layernorm.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.self_attn.k_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.self_attn.o_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.self_attn.q_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.96.self_attn.v_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.97.input_layernorm.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.97.mlp.down_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.97.mlp.gate_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.97.mlp.up_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.97.post_attention_layernorm.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.97.self_attn.k_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.97.self_attn.o_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.97.self_attn.q_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.97.self_attn.v_proj.weight": "pytorch_model-00035-of-00050.bin", + "model.layers.98.input_layernorm.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.mlp.down_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.mlp.gate_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.mlp.up_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.post_attention_layernorm.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.self_attn.k_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.self_attn.o_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.self_attn.q_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.98.self_attn.v_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.input_layernorm.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.mlp.down_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.mlp.gate_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.mlp.up_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.post_attention_layernorm.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.self_attn.k_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.self_attn.o_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.self_attn.q_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.layers.99.self_attn.v_proj.weight": "pytorch_model-00036-of-00050.bin", + "model.norm.weight": "pytorch_model-00049-of-00050.bin" + } +} diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..c32a4966272ecd7ef0b10de11e59a0541d4cf18b --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,35 @@ +{ + "additional_special_tokens": [ + "", + "", + "" + ], + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/tokenizer.model b/tokenizer.model new file mode 100644 index 0000000000000000000000000000000000000000..6c00c742ce03c627d6cd5b795984876fa49fa899 --- /dev/null +++ b/tokenizer.model @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347 +size 499723 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5da7cdebc2203daf955e3e6ad61e303a41601ef3 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,72 @@ +{ + "add_bos_token": true, + "add_eos_token": false, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "32000": { + "content": "<|system|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32001": { + "content": "<|user|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32002": { + "content": "<|model|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "", + "" + ], + "bos_token": "", + "clean_up_tokenization_spaces": false, + "eos_token": "", + "legacy": false, + "model_max_length": 1000000000000000019884624838656, + "pad_token": "", + "sp_model_kwargs": {}, + "spaces_between_special_tokens": false, + "tokenizer_class": "LlamaTokenizer", + "trust_remote_code": false, + "unk_token": "", + "use_default_system_prompt": true, + "use_fast": true +}