khazic commited on
Commit
3124a50
·
verified ·
1 Parent(s): 62f64d8

Upload folder using huggingface_hub

Browse files
llama3-8B-eagle3-128k-2hs-fcnorm-fullmha-epoch0/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLMEagle3"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 128000,
8
+ "draft_vocab_size": 128256,
9
+ "enable_2hs": true,
10
+ "enable_fc_norm": true,
11
+ "eos_token_id": 128001,
12
+ "head_dim": 128,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 4096,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 14336,
17
+ "max_position_embeddings": 131072,
18
+ "mlp_bias": false,
19
+ "model_type": "llama",
20
+ "num_attention_heads": 32,
21
+ "num_hidden_layers": 1,
22
+ "num_key_value_heads": 32,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-05,
26
+ "rope_scaling": {
27
+ "factor": 8.0,
28
+ "high_freq_factor": 4.0,
29
+ "low_freq_factor": 1.0,
30
+ "original_max_position_embeddings": 8192,
31
+ "rope_type": "llama3"
32
+ },
33
+ "rope_theta": 500000.0,
34
+ "tie_word_embeddings": false,
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.55.2",
37
+ "use_cache": true,
38
+ "vocab_size": 128256
39
+ }
llama3-8B-eagle3-128k-2hs-fcnorm-fullmha-epoch0/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e03f9d95c305c83830b7f37d6c53543529e448e8c0317d473455db3db1b3bb71
3
+ size 2907866088
llama3-8B-eagle3-128k-2hs-fcnorm-fullmha-epoch0/training_state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35843dd10d9399d4a1fbe630c92c6e6d64fea8362e3d0ba5c0c546b5706799a3
3
+ size 10263