Upload config.json with huggingface_hub
Browse files- config.json +51 -0
config.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"emb_dim": 256,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"DFlashDraftModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"auto_map": {
|
| 9 |
+
"AutoModel": "dflash.DFlashDraftModel"
|
| 10 |
+
},
|
| 11 |
+
"block_size": 16,
|
| 12 |
+
"bos_token_id": 151643,
|
| 13 |
+
"dflash_config": {
|
| 14 |
+
"mask_token_id": 151669,
|
| 15 |
+
"target_layer_ids": [1, 9, 17, 25, 33],
|
| 16 |
+
"projector_type": "domino",
|
| 17 |
+
"pure_draft_prefix_len": 1,
|
| 18 |
+
"emb_dim": 256,
|
| 19 |
+
"gru_hidden_dim": 1024,
|
| 20 |
+
"shift_label": true
|
| 21 |
+
},
|
| 22 |
+
"dtype": "bfloat16",
|
| 23 |
+
"eos_token_id": 151645,
|
| 24 |
+
"head_dim": 128,
|
| 25 |
+
"hidden_act": "silu",
|
| 26 |
+
"hidden_size": 4096,
|
| 27 |
+
"initializer_range": 0.02,
|
| 28 |
+
"intermediate_size": 12288,
|
| 29 |
+
"layer_types": [
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention"
|
| 35 |
+
],
|
| 36 |
+
"max_position_embeddings": 40960,
|
| 37 |
+
"max_window_layers": 5,
|
| 38 |
+
"model_type": "qwen3",
|
| 39 |
+
"num_attention_heads": 32,
|
| 40 |
+
"num_hidden_layers": 5,
|
| 41 |
+
"num_key_value_heads": 8,
|
| 42 |
+
"num_target_layers": 36,
|
| 43 |
+
"rms_norm_eps": 1e-06,
|
| 44 |
+
"rope_scaling": null,
|
| 45 |
+
"rope_theta": 1000000,
|
| 46 |
+
"sliding_window": null,
|
| 47 |
+
"tie_word_embeddings": false,
|
| 48 |
+
"use_cache": true,
|
| 49 |
+
"use_sliding_window": false,
|
| 50 |
+
"vocab_size": 151936
|
| 51 |
+
}
|