shahules786
commited on
Commit
•
3a6b70b
1
Parent(s):
b4f2ae2
Upload config.json with huggingface_hub
Browse files- config.json +7 -5
config.json
CHANGED
@@ -1,21 +1,21 @@
|
|
1 |
{
|
2 |
"_name_or_path": "t5-base",
|
3 |
"architectures": [
|
4 |
-
"
|
5 |
],
|
6 |
"d_ff": 3072,
|
7 |
"d_kv": 64,
|
8 |
"d_model": 768,
|
9 |
"decoder_start_token_id": 0,
|
10 |
-
"dense_act_fn": "
|
11 |
"dropout_rate": 0.1,
|
12 |
"eos_token_id": 1,
|
13 |
-
"feed_forward_proj": "
|
14 |
"initializer_factor": 1.0,
|
15 |
"is_encoder_decoder": true,
|
16 |
-
"is_gated_act":
|
17 |
"layer_norm_epsilon": 1e-06,
|
18 |
-
"model_type": "
|
19 |
"n_positions": 512,
|
20 |
"num_decoder_layers": 12,
|
21 |
"num_heads": 12,
|
@@ -53,6 +53,8 @@
|
|
53 |
"prefix": "translate English to Romanian: "
|
54 |
}
|
55 |
},
|
|
|
|
|
56 |
"torch_dtype": "float32",
|
57 |
"transformers_version": "4.26.1",
|
58 |
"use_cache": true,
|
|
|
1 |
{
|
2 |
"_name_or_path": "t5-base",
|
3 |
"architectures": [
|
4 |
+
"MT5ForConditionalGeneration"
|
5 |
],
|
6 |
"d_ff": 3072,
|
7 |
"d_kv": 64,
|
8 |
"d_model": 768,
|
9 |
"decoder_start_token_id": 0,
|
10 |
+
"dense_act_fn": "gelu_new",
|
11 |
"dropout_rate": 0.1,
|
12 |
"eos_token_id": 1,
|
13 |
+
"feed_forward_proj": "gated-gelu",
|
14 |
"initializer_factor": 1.0,
|
15 |
"is_encoder_decoder": true,
|
16 |
+
"is_gated_act": true,
|
17 |
"layer_norm_epsilon": 1e-06,
|
18 |
+
"model_type": "mt5",
|
19 |
"n_positions": 512,
|
20 |
"num_decoder_layers": 12,
|
21 |
"num_heads": 12,
|
|
|
53 |
"prefix": "translate English to Romanian: "
|
54 |
}
|
55 |
},
|
56 |
+
"tie_word_embeddings": false,
|
57 |
+
"tokenizer_class": "T5Tokenizer",
|
58 |
"torch_dtype": "float32",
|
59 |
"transformers_version": "4.26.1",
|
60 |
"use_cache": true,
|