Upload folder using huggingface_hub
Browse files
eagle3-128k-2hs-i2x-fullmha/epoch_0/config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"eos_token_id": 128001,
|
| 11 |
+
"head_dim": 128,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 4096,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 8192,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"num_attention_heads": 32,
|
| 20 |
+
"num_hidden_layers": 1,
|
| 21 |
+
"num_key_value_heads": 32,
|
| 22 |
+
"pad_token_id": 0,
|
| 23 |
+
"pretraining_tp": 1,
|
| 24 |
+
"rms_norm_eps": 1e-05,
|
| 25 |
+
"rope_scaling": {
|
| 26 |
+
"factor": 8.0,
|
| 27 |
+
"high_freq_factor": 4.0,
|
| 28 |
+
"low_freq_factor": 1.0,
|
| 29 |
+
"original_max_position_embeddings": 8192,
|
| 30 |
+
"rope_type": "llama3"
|
| 31 |
+
},
|
| 32 |
+
"rope_theta": 500000.0,
|
| 33 |
+
"tie_word_embeddings": false,
|
| 34 |
+
"torch_dtype": "bfloat16",
|
| 35 |
+
"transformers_version": "4.55.2",
|
| 36 |
+
"use_cache": true,
|
| 37 |
+
"vocab_size": 128256
|
| 38 |
+
}
|
eagle3-128k-2hs-i2x-fullmha/epoch_0/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65c934a8f11a6b5421c9fba80c3edeb8669cbcc004eea60fd210045dac09f5a4
|
| 3 |
+
size 2706514832
|
eagle3-128k-2hs-i2x-fullmha/epoch_0/training_state.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f08bee32089c7efdf1026e18da9ec5fca00ec09e0d25fe9ab92ad682ed8bd528
|
| 3 |
+
size 10263
|