epoch-02 (step 21528)
Browse files- epoch-02/config.json +62 -0
- epoch-02/model.safetensors +3 -0
epoch-02/config.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DSparkDraftModel"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"auto_map": {
|
| 8 |
+
"AutoModel": "dspark.DSparkDraftModel"
|
| 9 |
+
},
|
| 10 |
+
"block_size": 16,
|
| 11 |
+
"bos_token_id": 151643,
|
| 12 |
+
"dflash_config": {
|
| 13 |
+
"attention_mode": "gqa",
|
| 14 |
+
"confidence_head_alpha": 0.0,
|
| 15 |
+
"confidence_head_with_markov": false,
|
| 16 |
+
"enable_confidence_head": false,
|
| 17 |
+
"markov_head_type": "vanilla",
|
| 18 |
+
"markov_rank": 256,
|
| 19 |
+
"mask_token_id": 151669,
|
| 20 |
+
"projector_type": "dspark",
|
| 21 |
+
"target_layer_ids": [
|
| 22 |
+
1,
|
| 23 |
+
9,
|
| 24 |
+
17,
|
| 25 |
+
25,
|
| 26 |
+
33
|
| 27 |
+
]
|
| 28 |
+
},
|
| 29 |
+
"draft_vocab_size": 151936,
|
| 30 |
+
"dtype": "bfloat16",
|
| 31 |
+
"eos_token_id": 151645,
|
| 32 |
+
"head_dim": 128,
|
| 33 |
+
"hidden_act": "silu",
|
| 34 |
+
"hidden_size": 4096,
|
| 35 |
+
"initializer_range": 0.02,
|
| 36 |
+
"intermediate_size": 1536,
|
| 37 |
+
"is_causal": true,
|
| 38 |
+
"layer_types": [
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention"
|
| 41 |
+
],
|
| 42 |
+
"max_position_embeddings": 40960,
|
| 43 |
+
"max_window_layers": 2,
|
| 44 |
+
"model_type": "qwen3",
|
| 45 |
+
"num_attention_heads": 8,
|
| 46 |
+
"num_hidden_layers": 2,
|
| 47 |
+
"num_key_value_heads": 2,
|
| 48 |
+
"num_target_layers": 36,
|
| 49 |
+
"pad_token_id": null,
|
| 50 |
+
"rms_norm_eps": 1e-06,
|
| 51 |
+
"rope_parameters": {
|
| 52 |
+
"rope_theta": 1000000,
|
| 53 |
+
"rope_type": "default"
|
| 54 |
+
},
|
| 55 |
+
"rope_theta": 1000000,
|
| 56 |
+
"sliding_window": null,
|
| 57 |
+
"tie_word_embeddings": false,
|
| 58 |
+
"transformers_version": "5.12.1",
|
| 59 |
+
"use_cache": true,
|
| 60 |
+
"use_sliding_window": false,
|
| 61 |
+
"vocab_size": 151936
|
| 62 |
+
}
|
epoch-02/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2157ee8b2d7adba5e3e138a0c317b47e6ff90490627cf73cfc15f074c4537046
|
| 3 |
+
size 440848160
|