mlc-chat-config.json

{
  "model_type": "rwkv5",
  "quantization": "q4f16_2",
  "model_config": {
    "hidden_size": 2560,
    "intermediate_size": 8960,
    "num_hidden_layers": 32,
    "vocab_size": 65536,
    "model_version": "5_2",
    "tensor_parallel_shards": 1,
    "rescale_every": 6,
    "head_size": 64,
    "layer_norm_epsilon": 1e-05,
    "context_window_size": -1,
    "prefill_chunk_size": 4096,
    "num_heads": 40
  },
  "vocab_size": 65536,
  "context_window_size": -1,
  "sliding_window_size": -1,
  "prefill_chunk_size": 4096,
  "attention_sink_size": -1,
  "tensor_parallel_shards": 1,
  "mean_gen_len": 128,
  "max_gen_len": 4096,
  "shift_fill_factor": 0.3,
  "temperature": 0.7,
  "repetition_penalty": 1.0,
  "top_p": 0.95,
  "conv_template": "rwkv_world",
  "pad_token_id": 0,
  "bos_token_id": 0,
  "eos_token_id": 0,
  "tokenizer_files": [
    "tokenizer_model"
  ],
  "version": "0.1.0"
}