Upload config.json with huggingface_hub
Browse files- config.json +89 -0
config.json
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_classification_head": false,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"HunYuanDenseV1ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attention_head_dim": 128,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"cla_share_factor": 2,
|
| 11 |
+
"class_num": 0,
|
| 12 |
+
"dense_list": [
|
| 13 |
+
4096,
|
| 14 |
+
0
|
| 15 |
+
],
|
| 16 |
+
"eos_token_id": 127960,
|
| 17 |
+
"head_dim": 128,
|
| 18 |
+
"hidden_act": "silu",
|
| 19 |
+
"hidden_size": 4096,
|
| 20 |
+
"im_end_id": 5,
|
| 21 |
+
"im_newline_id": 11,
|
| 22 |
+
"im_start_id": 4,
|
| 23 |
+
"initializer_range": 0.02,
|
| 24 |
+
"intermediate_size": 14336,
|
| 25 |
+
"mask_init_id": 12,
|
| 26 |
+
"max_position_embeddings": 10240,
|
| 27 |
+
"mlp_bias": false,
|
| 28 |
+
"model_type": "hunyuan_v1_dense",
|
| 29 |
+
"norm_type": "rms",
|
| 30 |
+
"num_attention_heads": 32,
|
| 31 |
+
"num_hidden_layers": 32,
|
| 32 |
+
"num_key_value_heads": 8,
|
| 33 |
+
"org_vocab_size": 290943,
|
| 34 |
+
"pad_id": 127961,
|
| 35 |
+
"pad_token_id": 0,
|
| 36 |
+
"pool_type": "last",
|
| 37 |
+
"pretraining_tp": 1,
|
| 38 |
+
"quantization_config": {
|
| 39 |
+
"config_groups": {
|
| 40 |
+
"group_0": {
|
| 41 |
+
"input_activations": {
|
| 42 |
+
"dynamic": false,
|
| 43 |
+
"num_bits": 8,
|
| 44 |
+
"strategy": "tensor",
|
| 45 |
+
"type": "float"
|
| 46 |
+
},
|
| 47 |
+
"output_activations": null,
|
| 48 |
+
"targets": [
|
| 49 |
+
"Linear"
|
| 50 |
+
],
|
| 51 |
+
"weights": {
|
| 52 |
+
"dynamic": false,
|
| 53 |
+
"num_bits": 8,
|
| 54 |
+
"strategy": "tensor",
|
| 55 |
+
"type": "float"
|
| 56 |
+
}
|
| 57 |
+
}
|
| 58 |
+
},
|
| 59 |
+
"format": "naive-quantized",
|
| 60 |
+
"ignored_layers": [
|
| 61 |
+
"lm_head",
|
| 62 |
+
"model.embed_tokens"
|
| 63 |
+
],
|
| 64 |
+
"kv_cache_scheme": null,
|
| 65 |
+
"quant_method": "compressed-tensors",
|
| 66 |
+
"quantization_status": "compressed"
|
| 67 |
+
},
|
| 68 |
+
"rms_norm_eps": 1e-05,
|
| 69 |
+
"rope_scaling": {
|
| 70 |
+
"alpha": 100000.0,
|
| 71 |
+
"beta_fast": 32,
|
| 72 |
+
"beta_slow": 1,
|
| 73 |
+
"factor": 1.0,
|
| 74 |
+
"mscale": 1.0,
|
| 75 |
+
"mscale_all_dim": 1.0,
|
| 76 |
+
"type": "dynamic"
|
| 77 |
+
},
|
| 78 |
+
"rope_theta": 10000.0,
|
| 79 |
+
"text_end_id": 7,
|
| 80 |
+
"text_start_id": 6,
|
| 81 |
+
"tie_word_embeddings": true,
|
| 82 |
+
"torch_dtype": "bfloat16",
|
| 83 |
+
"transformers_version": "4.55.0.dev0",
|
| 84 |
+
"use_cache": false,
|
| 85 |
+
"use_cla": false,
|
| 86 |
+
"use_qk_norm": true,
|
| 87 |
+
"use_rotary_pos_emb": true,
|
| 88 |
+
"vocab_size": 128256
|
| 89 |
+
}
|