Chien commited on
Upload config.json with huggingface_hub
Browse files- config.json +56 -0
config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 3 |
+
"activation_type": "silu",
|
| 4 |
+
"alibi": false,
|
| 5 |
+
"alibi_bias_max": 8.0,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"LLaDAModelLM"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"attention_layer_norm": false,
|
| 11 |
+
"attention_layer_norm_with_affine": true,
|
| 12 |
+
"auto_map": {
|
| 13 |
+
"AutoConfig": "GSAI-ML/LLaDA-8B-Instruct--configuration_llada.LLaDAConfig",
|
| 14 |
+
"AutoModel": "GSAI-ML/LLaDA-8B-Instruct--modeling_llada.LLaDAModelLM",
|
| 15 |
+
"AutoModelForCausalLM": "GSAI-ML/LLaDA-8B-Instruct--modeling_llada.LLaDAModelLM"
|
| 16 |
+
},
|
| 17 |
+
"bias_for_layer_norm": false,
|
| 18 |
+
"block_group_size": 1,
|
| 19 |
+
"block_type": "llama",
|
| 20 |
+
"d_model": 4096,
|
| 21 |
+
"embedding_dropout": 0.0,
|
| 22 |
+
"embedding_size": 126464,
|
| 23 |
+
"eos_token_id": 126081,
|
| 24 |
+
"flash_attention": false,
|
| 25 |
+
"include_bias": false,
|
| 26 |
+
"include_qkv_bias": false,
|
| 27 |
+
"init_cutoff_factor": null,
|
| 28 |
+
"init_device": "meta",
|
| 29 |
+
"init_fn": "mitchell",
|
| 30 |
+
"init_std": 0.02,
|
| 31 |
+
"input_emb_norm": false,
|
| 32 |
+
"layer_norm_type": "rms",
|
| 33 |
+
"layer_norm_with_affine": true,
|
| 34 |
+
"mask_token_id": 126336,
|
| 35 |
+
"max_sequence_length": 4096,
|
| 36 |
+
"mlp_hidden_size": 12288,
|
| 37 |
+
"mlp_ratio": 4,
|
| 38 |
+
"model_type": "llada",
|
| 39 |
+
"multi_query_attention": null,
|
| 40 |
+
"n_heads": 32,
|
| 41 |
+
"n_kv_heads": 32,
|
| 42 |
+
"n_layers": 32,
|
| 43 |
+
"pad_token_id": 126081,
|
| 44 |
+
"precision": "amp_bf16",
|
| 45 |
+
"residual_dropout": 0.0,
|
| 46 |
+
"rms_norm_eps": 1e-05,
|
| 47 |
+
"rope": true,
|
| 48 |
+
"rope_full_precision": true,
|
| 49 |
+
"rope_theta": 500000.0,
|
| 50 |
+
"scale_logits": false,
|
| 51 |
+
"torch_dtype": "bfloat16",
|
| 52 |
+
"transformers_version": "4.49.0",
|
| 53 |
+
"use_cache": false,
|
| 54 |
+
"vocab_size": 126464,
|
| 55 |
+
"weight_tying": false
|
| 56 |
+
}
|