Commit
•
ceb1847
1
Parent(s):
80785bb
Training in progress, step 2000, checkpoint
Browse files- checkpoint-2000/config.json +32 -0
- checkpoint-2000/global_step2000/mp_rank_00_model_states.pt +3 -0
- checkpoint-2000/global_step2000/zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- checkpoint-2000/latest +1 -0
- checkpoint-2000/model-00001-of-00002.safetensors +3 -0
- checkpoint-2000/model-00002-of-00002.safetensors +3 -0
- checkpoint-2000/model.safetensors.index.json +396 -0
- checkpoint-2000/rng_state_0.pth +3 -0
- checkpoint-2000/scheduler.pt +3 -0
- checkpoint-2000/special_tokens_map.json +24 -0
- checkpoint-2000/tokenizer.json +0 -0
- checkpoint-2000/tokenizer_config.json +214 -0
- checkpoint-2000/trainer_state.json +1485 -0
- checkpoint-2000/training_args.bin +3 -0
- checkpoint-2000/zero_to_fp32.py +604 -0
checkpoint-2000/config.json
ADDED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "EleutherAI/pythia-2.8b",
|
3 |
+
"architectures": [
|
4 |
+
"GPTNeoForMultipleChoice"
|
5 |
+
],
|
6 |
+
"attention_bias": true,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 0,
|
9 |
+
"classifier_dropout": 0.1,
|
10 |
+
"eos_token_id": 0,
|
11 |
+
"hidden_act": "gelu",
|
12 |
+
"hidden_dropout": 0.0,
|
13 |
+
"hidden_size": 2560,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 10240,
|
16 |
+
"layer_norm_eps": 1e-05,
|
17 |
+
"max_length": 1024,
|
18 |
+
"max_position_embeddings": 2048,
|
19 |
+
"model_type": "gpt_neox",
|
20 |
+
"num_attention_heads": 32,
|
21 |
+
"num_hidden_layers": 32,
|
22 |
+
"pad_token_id": 0,
|
23 |
+
"rope_scaling": null,
|
24 |
+
"rotary_emb_base": 10000,
|
25 |
+
"rotary_pct": 0.25,
|
26 |
+
"tie_word_embeddings": false,
|
27 |
+
"torch_dtype": "float16",
|
28 |
+
"transformers_version": "4.41.0",
|
29 |
+
"use_cache": true,
|
30 |
+
"use_parallel_residual": true,
|
31 |
+
"vocab_size": 50304
|
32 |
+
}
|
checkpoint-2000/global_step2000/mp_rank_00_model_states.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b05cd061d2aaea80e2ef00c861261a83f7b26e0c0eba9ab26bcc8d697892d80a
|
3 |
+
size 5292979768
|
checkpoint-2000/global_step2000/zero_pp_rank_0_mp_rank_00_optim_states.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:596b3a2505936459789bbf097685251abae20b53e2df4f792be0ce31e43e048e
|
3 |
+
size 15878620944
|
checkpoint-2000/latest
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
global_step2000
|
checkpoint-2000/model-00001-of-00002.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:063a5c51639e3d9b7ed68cda60cf027da588ed9c90bc8a3c302eec5a9bbf575f
|
3 |
+
size 4978208880
|
checkpoint-2000/model-00002-of-00002.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b1c30f212b36eb542087b33821869d6c9e615f2bfaf492bc8fea09f27e420405
|
3 |
+
size 314703498
|
checkpoint-2000/model.safetensors.index.json
ADDED
@@ -0,0 +1,396 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"metadata": {
|
3 |
+
"total_size": 5292866562
|
4 |
+
},
|
5 |
+
"weight_map": {
|
6 |
+
"classifier.bias": "model-00002-of-00002.safetensors",
|
7 |
+
"classifier.weight": "model-00002-of-00002.safetensors",
|
8 |
+
"gpt_neox.embed_in.weight": "model-00001-of-00002.safetensors",
|
9 |
+
"gpt_neox.final_layer_norm.bias": "model-00002-of-00002.safetensors",
|
10 |
+
"gpt_neox.final_layer_norm.weight": "model-00002-of-00002.safetensors",
|
11 |
+
"gpt_neox.layers.0.attention.dense.bias": "model-00001-of-00002.safetensors",
|
12 |
+
"gpt_neox.layers.0.attention.dense.weight": "model-00001-of-00002.safetensors",
|
13 |
+
"gpt_neox.layers.0.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
14 |
+
"gpt_neox.layers.0.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
15 |
+
"gpt_neox.layers.0.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
16 |
+
"gpt_neox.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
17 |
+
"gpt_neox.layers.0.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
18 |
+
"gpt_neox.layers.0.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
19 |
+
"gpt_neox.layers.0.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
20 |
+
"gpt_neox.layers.0.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
21 |
+
"gpt_neox.layers.0.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
22 |
+
"gpt_neox.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
23 |
+
"gpt_neox.layers.1.attention.dense.bias": "model-00001-of-00002.safetensors",
|
24 |
+
"gpt_neox.layers.1.attention.dense.weight": "model-00001-of-00002.safetensors",
|
25 |
+
"gpt_neox.layers.1.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
26 |
+
"gpt_neox.layers.1.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
27 |
+
"gpt_neox.layers.1.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
28 |
+
"gpt_neox.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
29 |
+
"gpt_neox.layers.1.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
30 |
+
"gpt_neox.layers.1.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
31 |
+
"gpt_neox.layers.1.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
32 |
+
"gpt_neox.layers.1.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
33 |
+
"gpt_neox.layers.1.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
34 |
+
"gpt_neox.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
35 |
+
"gpt_neox.layers.10.attention.dense.bias": "model-00001-of-00002.safetensors",
|
36 |
+
"gpt_neox.layers.10.attention.dense.weight": "model-00001-of-00002.safetensors",
|
37 |
+
"gpt_neox.layers.10.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
38 |
+
"gpt_neox.layers.10.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
39 |
+
"gpt_neox.layers.10.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
40 |
+
"gpt_neox.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
41 |
+
"gpt_neox.layers.10.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
42 |
+
"gpt_neox.layers.10.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
43 |
+
"gpt_neox.layers.10.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
44 |
+
"gpt_neox.layers.10.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
45 |
+
"gpt_neox.layers.10.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
46 |
+
"gpt_neox.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
47 |
+
"gpt_neox.layers.11.attention.dense.bias": "model-00001-of-00002.safetensors",
|
48 |
+
"gpt_neox.layers.11.attention.dense.weight": "model-00001-of-00002.safetensors",
|
49 |
+
"gpt_neox.layers.11.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
50 |
+
"gpt_neox.layers.11.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
51 |
+
"gpt_neox.layers.11.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
52 |
+
"gpt_neox.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
53 |
+
"gpt_neox.layers.11.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
54 |
+
"gpt_neox.layers.11.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
55 |
+
"gpt_neox.layers.11.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
56 |
+
"gpt_neox.layers.11.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
57 |
+
"gpt_neox.layers.11.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
58 |
+
"gpt_neox.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
59 |
+
"gpt_neox.layers.12.attention.dense.bias": "model-00001-of-00002.safetensors",
|
60 |
+
"gpt_neox.layers.12.attention.dense.weight": "model-00001-of-00002.safetensors",
|
61 |
+
"gpt_neox.layers.12.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
62 |
+
"gpt_neox.layers.12.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
63 |
+
"gpt_neox.layers.12.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
64 |
+
"gpt_neox.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
65 |
+
"gpt_neox.layers.12.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
66 |
+
"gpt_neox.layers.12.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
67 |
+
"gpt_neox.layers.12.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
68 |
+
"gpt_neox.layers.12.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
69 |
+
"gpt_neox.layers.12.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
70 |
+
"gpt_neox.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
71 |
+
"gpt_neox.layers.13.attention.dense.bias": "model-00001-of-00002.safetensors",
|
72 |
+
"gpt_neox.layers.13.attention.dense.weight": "model-00001-of-00002.safetensors",
|
73 |
+
"gpt_neox.layers.13.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
74 |
+
"gpt_neox.layers.13.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
75 |
+
"gpt_neox.layers.13.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
76 |
+
"gpt_neox.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
77 |
+
"gpt_neox.layers.13.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
78 |
+
"gpt_neox.layers.13.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
79 |
+
"gpt_neox.layers.13.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
80 |
+
"gpt_neox.layers.13.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
81 |
+
"gpt_neox.layers.13.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
82 |
+
"gpt_neox.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
83 |
+
"gpt_neox.layers.14.attention.dense.bias": "model-00001-of-00002.safetensors",
|
84 |
+
"gpt_neox.layers.14.attention.dense.weight": "model-00001-of-00002.safetensors",
|
85 |
+
"gpt_neox.layers.14.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
86 |
+
"gpt_neox.layers.14.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
87 |
+
"gpt_neox.layers.14.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
88 |
+
"gpt_neox.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
89 |
+
"gpt_neox.layers.14.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
90 |
+
"gpt_neox.layers.14.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
91 |
+
"gpt_neox.layers.14.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
92 |
+
"gpt_neox.layers.14.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
93 |
+
"gpt_neox.layers.14.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
94 |
+
"gpt_neox.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
95 |
+
"gpt_neox.layers.15.attention.dense.bias": "model-00001-of-00002.safetensors",
|
96 |
+
"gpt_neox.layers.15.attention.dense.weight": "model-00001-of-00002.safetensors",
|
97 |
+
"gpt_neox.layers.15.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
98 |
+
"gpt_neox.layers.15.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
99 |
+
"gpt_neox.layers.15.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
100 |
+
"gpt_neox.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
101 |
+
"gpt_neox.layers.15.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
102 |
+
"gpt_neox.layers.15.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
103 |
+
"gpt_neox.layers.15.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
104 |
+
"gpt_neox.layers.15.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
105 |
+
"gpt_neox.layers.15.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
106 |
+
"gpt_neox.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
107 |
+
"gpt_neox.layers.16.attention.dense.bias": "model-00001-of-00002.safetensors",
|
108 |
+
"gpt_neox.layers.16.attention.dense.weight": "model-00001-of-00002.safetensors",
|
109 |
+
"gpt_neox.layers.16.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
110 |
+
"gpt_neox.layers.16.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
111 |
+
"gpt_neox.layers.16.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
112 |
+
"gpt_neox.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
113 |
+
"gpt_neox.layers.16.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
114 |
+
"gpt_neox.layers.16.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
115 |
+
"gpt_neox.layers.16.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
116 |
+
"gpt_neox.layers.16.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
117 |
+
"gpt_neox.layers.16.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
118 |
+
"gpt_neox.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
119 |
+
"gpt_neox.layers.17.attention.dense.bias": "model-00001-of-00002.safetensors",
|
120 |
+
"gpt_neox.layers.17.attention.dense.weight": "model-00001-of-00002.safetensors",
|
121 |
+
"gpt_neox.layers.17.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
122 |
+
"gpt_neox.layers.17.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
123 |
+
"gpt_neox.layers.17.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
124 |
+
"gpt_neox.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
125 |
+
"gpt_neox.layers.17.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
126 |
+
"gpt_neox.layers.17.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
127 |
+
"gpt_neox.layers.17.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
128 |
+
"gpt_neox.layers.17.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
129 |
+
"gpt_neox.layers.17.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
130 |
+
"gpt_neox.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
131 |
+
"gpt_neox.layers.18.attention.dense.bias": "model-00001-of-00002.safetensors",
|
132 |
+
"gpt_neox.layers.18.attention.dense.weight": "model-00001-of-00002.safetensors",
|
133 |
+
"gpt_neox.layers.18.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
134 |
+
"gpt_neox.layers.18.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
135 |
+
"gpt_neox.layers.18.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
136 |
+
"gpt_neox.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
137 |
+
"gpt_neox.layers.18.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
138 |
+
"gpt_neox.layers.18.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
139 |
+
"gpt_neox.layers.18.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
140 |
+
"gpt_neox.layers.18.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
141 |
+
"gpt_neox.layers.18.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
142 |
+
"gpt_neox.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
143 |
+
"gpt_neox.layers.19.attention.dense.bias": "model-00001-of-00002.safetensors",
|
144 |
+
"gpt_neox.layers.19.attention.dense.weight": "model-00001-of-00002.safetensors",
|
145 |
+
"gpt_neox.layers.19.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
146 |
+
"gpt_neox.layers.19.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
147 |
+
"gpt_neox.layers.19.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
148 |
+
"gpt_neox.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
149 |
+
"gpt_neox.layers.19.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
150 |
+
"gpt_neox.layers.19.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
151 |
+
"gpt_neox.layers.19.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
152 |
+
"gpt_neox.layers.19.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
153 |
+
"gpt_neox.layers.19.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
154 |
+
"gpt_neox.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
155 |
+
"gpt_neox.layers.2.attention.dense.bias": "model-00001-of-00002.safetensors",
|
156 |
+
"gpt_neox.layers.2.attention.dense.weight": "model-00001-of-00002.safetensors",
|
157 |
+
"gpt_neox.layers.2.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
158 |
+
"gpt_neox.layers.2.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
159 |
+
"gpt_neox.layers.2.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
160 |
+
"gpt_neox.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
161 |
+
"gpt_neox.layers.2.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
162 |
+
"gpt_neox.layers.2.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
163 |
+
"gpt_neox.layers.2.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
164 |
+
"gpt_neox.layers.2.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
165 |
+
"gpt_neox.layers.2.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
166 |
+
"gpt_neox.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
167 |
+
"gpt_neox.layers.20.attention.dense.bias": "model-00001-of-00002.safetensors",
|
168 |
+
"gpt_neox.layers.20.attention.dense.weight": "model-00001-of-00002.safetensors",
|
169 |
+
"gpt_neox.layers.20.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
170 |
+
"gpt_neox.layers.20.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
171 |
+
"gpt_neox.layers.20.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
172 |
+
"gpt_neox.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
173 |
+
"gpt_neox.layers.20.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
174 |
+
"gpt_neox.layers.20.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
175 |
+
"gpt_neox.layers.20.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
176 |
+
"gpt_neox.layers.20.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
177 |
+
"gpt_neox.layers.20.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
178 |
+
"gpt_neox.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
179 |
+
"gpt_neox.layers.21.attention.dense.bias": "model-00001-of-00002.safetensors",
|
180 |
+
"gpt_neox.layers.21.attention.dense.weight": "model-00001-of-00002.safetensors",
|
181 |
+
"gpt_neox.layers.21.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
182 |
+
"gpt_neox.layers.21.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
183 |
+
"gpt_neox.layers.21.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
184 |
+
"gpt_neox.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
185 |
+
"gpt_neox.layers.21.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
186 |
+
"gpt_neox.layers.21.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
187 |
+
"gpt_neox.layers.21.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
188 |
+
"gpt_neox.layers.21.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
189 |
+
"gpt_neox.layers.21.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
190 |
+
"gpt_neox.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
191 |
+
"gpt_neox.layers.22.attention.dense.bias": "model-00001-of-00002.safetensors",
|
192 |
+
"gpt_neox.layers.22.attention.dense.weight": "model-00001-of-00002.safetensors",
|
193 |
+
"gpt_neox.layers.22.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
194 |
+
"gpt_neox.layers.22.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
195 |
+
"gpt_neox.layers.22.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
196 |
+
"gpt_neox.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
197 |
+
"gpt_neox.layers.22.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
198 |
+
"gpt_neox.layers.22.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
199 |
+
"gpt_neox.layers.22.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
200 |
+
"gpt_neox.layers.22.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
201 |
+
"gpt_neox.layers.22.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
202 |
+
"gpt_neox.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
203 |
+
"gpt_neox.layers.23.attention.dense.bias": "model-00001-of-00002.safetensors",
|
204 |
+
"gpt_neox.layers.23.attention.dense.weight": "model-00001-of-00002.safetensors",
|
205 |
+
"gpt_neox.layers.23.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
206 |
+
"gpt_neox.layers.23.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
207 |
+
"gpt_neox.layers.23.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
208 |
+
"gpt_neox.layers.23.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
209 |
+
"gpt_neox.layers.23.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
210 |
+
"gpt_neox.layers.23.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
211 |
+
"gpt_neox.layers.23.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
212 |
+
"gpt_neox.layers.23.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
213 |
+
"gpt_neox.layers.23.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
214 |
+
"gpt_neox.layers.23.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
215 |
+
"gpt_neox.layers.24.attention.dense.bias": "model-00001-of-00002.safetensors",
|
216 |
+
"gpt_neox.layers.24.attention.dense.weight": "model-00001-of-00002.safetensors",
|
217 |
+
"gpt_neox.layers.24.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
218 |
+
"gpt_neox.layers.24.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
219 |
+
"gpt_neox.layers.24.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
220 |
+
"gpt_neox.layers.24.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
221 |
+
"gpt_neox.layers.24.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
222 |
+
"gpt_neox.layers.24.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
223 |
+
"gpt_neox.layers.24.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
224 |
+
"gpt_neox.layers.24.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
225 |
+
"gpt_neox.layers.24.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
226 |
+
"gpt_neox.layers.24.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
227 |
+
"gpt_neox.layers.25.attention.dense.bias": "model-00001-of-00002.safetensors",
|
228 |
+
"gpt_neox.layers.25.attention.dense.weight": "model-00001-of-00002.safetensors",
|
229 |
+
"gpt_neox.layers.25.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
230 |
+
"gpt_neox.layers.25.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
231 |
+
"gpt_neox.layers.25.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
232 |
+
"gpt_neox.layers.25.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
233 |
+
"gpt_neox.layers.25.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
234 |
+
"gpt_neox.layers.25.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
235 |
+
"gpt_neox.layers.25.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
236 |
+
"gpt_neox.layers.25.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
237 |
+
"gpt_neox.layers.25.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
238 |
+
"gpt_neox.layers.25.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
239 |
+
"gpt_neox.layers.26.attention.dense.bias": "model-00001-of-00002.safetensors",
|
240 |
+
"gpt_neox.layers.26.attention.dense.weight": "model-00001-of-00002.safetensors",
|
241 |
+
"gpt_neox.layers.26.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
242 |
+
"gpt_neox.layers.26.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
243 |
+
"gpt_neox.layers.26.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
244 |
+
"gpt_neox.layers.26.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
245 |
+
"gpt_neox.layers.26.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
246 |
+
"gpt_neox.layers.26.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
247 |
+
"gpt_neox.layers.26.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
248 |
+
"gpt_neox.layers.26.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
249 |
+
"gpt_neox.layers.26.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
250 |
+
"gpt_neox.layers.26.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
251 |
+
"gpt_neox.layers.27.attention.dense.bias": "model-00001-of-00002.safetensors",
|
252 |
+
"gpt_neox.layers.27.attention.dense.weight": "model-00001-of-00002.safetensors",
|
253 |
+
"gpt_neox.layers.27.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
254 |
+
"gpt_neox.layers.27.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
255 |
+
"gpt_neox.layers.27.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
256 |
+
"gpt_neox.layers.27.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
257 |
+
"gpt_neox.layers.27.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
258 |
+
"gpt_neox.layers.27.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
259 |
+
"gpt_neox.layers.27.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
260 |
+
"gpt_neox.layers.27.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
261 |
+
"gpt_neox.layers.27.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
262 |
+
"gpt_neox.layers.27.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
263 |
+
"gpt_neox.layers.28.attention.dense.bias": "model-00001-of-00002.safetensors",
|
264 |
+
"gpt_neox.layers.28.attention.dense.weight": "model-00001-of-00002.safetensors",
|
265 |
+
"gpt_neox.layers.28.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
266 |
+
"gpt_neox.layers.28.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
267 |
+
"gpt_neox.layers.28.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
268 |
+
"gpt_neox.layers.28.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
269 |
+
"gpt_neox.layers.28.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
270 |
+
"gpt_neox.layers.28.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
271 |
+
"gpt_neox.layers.28.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
272 |
+
"gpt_neox.layers.28.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
273 |
+
"gpt_neox.layers.28.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
274 |
+
"gpt_neox.layers.28.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
275 |
+
"gpt_neox.layers.29.attention.dense.bias": "model-00001-of-00002.safetensors",
|
276 |
+
"gpt_neox.layers.29.attention.dense.weight": "model-00001-of-00002.safetensors",
|
277 |
+
"gpt_neox.layers.29.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
278 |
+
"gpt_neox.layers.29.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
279 |
+
"gpt_neox.layers.29.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
280 |
+
"gpt_neox.layers.29.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
281 |
+
"gpt_neox.layers.29.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
282 |
+
"gpt_neox.layers.29.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
283 |
+
"gpt_neox.layers.29.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
284 |
+
"gpt_neox.layers.29.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
285 |
+
"gpt_neox.layers.29.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
286 |
+
"gpt_neox.layers.29.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
287 |
+
"gpt_neox.layers.3.attention.dense.bias": "model-00001-of-00002.safetensors",
|
288 |
+
"gpt_neox.layers.3.attention.dense.weight": "model-00001-of-00002.safetensors",
|
289 |
+
"gpt_neox.layers.3.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
290 |
+
"gpt_neox.layers.3.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
291 |
+
"gpt_neox.layers.3.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
292 |
+
"gpt_neox.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
293 |
+
"gpt_neox.layers.3.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
294 |
+
"gpt_neox.layers.3.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
295 |
+
"gpt_neox.layers.3.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
296 |
+
"gpt_neox.layers.3.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
297 |
+
"gpt_neox.layers.3.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
298 |
+
"gpt_neox.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
299 |
+
"gpt_neox.layers.30.attention.dense.bias": "model-00002-of-00002.safetensors",
|
300 |
+
"gpt_neox.layers.30.attention.dense.weight": "model-00002-of-00002.safetensors",
|
301 |
+
"gpt_neox.layers.30.attention.query_key_value.bias": "model-00002-of-00002.safetensors",
|
302 |
+
"gpt_neox.layers.30.attention.query_key_value.weight": "model-00002-of-00002.safetensors",
|
303 |
+
"gpt_neox.layers.30.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
304 |
+
"gpt_neox.layers.30.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
305 |
+
"gpt_neox.layers.30.mlp.dense_4h_to_h.bias": "model-00002-of-00002.safetensors",
|
306 |
+
"gpt_neox.layers.30.mlp.dense_4h_to_h.weight": "model-00002-of-00002.safetensors",
|
307 |
+
"gpt_neox.layers.30.mlp.dense_h_to_4h.bias": "model-00002-of-00002.safetensors",
|
308 |
+
"gpt_neox.layers.30.mlp.dense_h_to_4h.weight": "model-00002-of-00002.safetensors",
|
309 |
+
"gpt_neox.layers.30.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
310 |
+
"gpt_neox.layers.30.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
311 |
+
"gpt_neox.layers.31.attention.dense.bias": "model-00002-of-00002.safetensors",
|
312 |
+
"gpt_neox.layers.31.attention.dense.weight": "model-00002-of-00002.safetensors",
|
313 |
+
"gpt_neox.layers.31.attention.query_key_value.bias": "model-00002-of-00002.safetensors",
|
314 |
+
"gpt_neox.layers.31.attention.query_key_value.weight": "model-00002-of-00002.safetensors",
|
315 |
+
"gpt_neox.layers.31.input_layernorm.bias": "model-00002-of-00002.safetensors",
|
316 |
+
"gpt_neox.layers.31.input_layernorm.weight": "model-00002-of-00002.safetensors",
|
317 |
+
"gpt_neox.layers.31.mlp.dense_4h_to_h.bias": "model-00002-of-00002.safetensors",
|
318 |
+
"gpt_neox.layers.31.mlp.dense_4h_to_h.weight": "model-00002-of-00002.safetensors",
|
319 |
+
"gpt_neox.layers.31.mlp.dense_h_to_4h.bias": "model-00002-of-00002.safetensors",
|
320 |
+
"gpt_neox.layers.31.mlp.dense_h_to_4h.weight": "model-00002-of-00002.safetensors",
|
321 |
+
"gpt_neox.layers.31.post_attention_layernorm.bias": "model-00002-of-00002.safetensors",
|
322 |
+
"gpt_neox.layers.31.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
|
323 |
+
"gpt_neox.layers.4.attention.dense.bias": "model-00001-of-00002.safetensors",
|
324 |
+
"gpt_neox.layers.4.attention.dense.weight": "model-00001-of-00002.safetensors",
|
325 |
+
"gpt_neox.layers.4.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
326 |
+
"gpt_neox.layers.4.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
327 |
+
"gpt_neox.layers.4.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
328 |
+
"gpt_neox.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
329 |
+
"gpt_neox.layers.4.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
330 |
+
"gpt_neox.layers.4.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
331 |
+
"gpt_neox.layers.4.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
332 |
+
"gpt_neox.layers.4.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
333 |
+
"gpt_neox.layers.4.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
334 |
+
"gpt_neox.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
335 |
+
"gpt_neox.layers.5.attention.dense.bias": "model-00001-of-00002.safetensors",
|
336 |
+
"gpt_neox.layers.5.attention.dense.weight": "model-00001-of-00002.safetensors",
|
337 |
+
"gpt_neox.layers.5.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
338 |
+
"gpt_neox.layers.5.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
339 |
+
"gpt_neox.layers.5.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
340 |
+
"gpt_neox.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
341 |
+
"gpt_neox.layers.5.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
342 |
+
"gpt_neox.layers.5.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
343 |
+
"gpt_neox.layers.5.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
344 |
+
"gpt_neox.layers.5.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
345 |
+
"gpt_neox.layers.5.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
346 |
+
"gpt_neox.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
347 |
+
"gpt_neox.layers.6.attention.dense.bias": "model-00001-of-00002.safetensors",
|
348 |
+
"gpt_neox.layers.6.attention.dense.weight": "model-00001-of-00002.safetensors",
|
349 |
+
"gpt_neox.layers.6.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
350 |
+
"gpt_neox.layers.6.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
351 |
+
"gpt_neox.layers.6.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
352 |
+
"gpt_neox.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
353 |
+
"gpt_neox.layers.6.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
354 |
+
"gpt_neox.layers.6.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
355 |
+
"gpt_neox.layers.6.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
356 |
+
"gpt_neox.layers.6.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
357 |
+
"gpt_neox.layers.6.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
358 |
+
"gpt_neox.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
359 |
+
"gpt_neox.layers.7.attention.dense.bias": "model-00001-of-00002.safetensors",
|
360 |
+
"gpt_neox.layers.7.attention.dense.weight": "model-00001-of-00002.safetensors",
|
361 |
+
"gpt_neox.layers.7.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
362 |
+
"gpt_neox.layers.7.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
363 |
+
"gpt_neox.layers.7.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
364 |
+
"gpt_neox.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
365 |
+
"gpt_neox.layers.7.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
366 |
+
"gpt_neox.layers.7.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
367 |
+
"gpt_neox.layers.7.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
368 |
+
"gpt_neox.layers.7.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
369 |
+
"gpt_neox.layers.7.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
370 |
+
"gpt_neox.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
371 |
+
"gpt_neox.layers.8.attention.dense.bias": "model-00001-of-00002.safetensors",
|
372 |
+
"gpt_neox.layers.8.attention.dense.weight": "model-00001-of-00002.safetensors",
|
373 |
+
"gpt_neox.layers.8.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
374 |
+
"gpt_neox.layers.8.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
375 |
+
"gpt_neox.layers.8.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
376 |
+
"gpt_neox.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
377 |
+
"gpt_neox.layers.8.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
378 |
+
"gpt_neox.layers.8.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
379 |
+
"gpt_neox.layers.8.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
380 |
+
"gpt_neox.layers.8.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
381 |
+
"gpt_neox.layers.8.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
382 |
+
"gpt_neox.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
|
383 |
+
"gpt_neox.layers.9.attention.dense.bias": "model-00001-of-00002.safetensors",
|
384 |
+
"gpt_neox.layers.9.attention.dense.weight": "model-00001-of-00002.safetensors",
|
385 |
+
"gpt_neox.layers.9.attention.query_key_value.bias": "model-00001-of-00002.safetensors",
|
386 |
+
"gpt_neox.layers.9.attention.query_key_value.weight": "model-00001-of-00002.safetensors",
|
387 |
+
"gpt_neox.layers.9.input_layernorm.bias": "model-00001-of-00002.safetensors",
|
388 |
+
"gpt_neox.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
|
389 |
+
"gpt_neox.layers.9.mlp.dense_4h_to_h.bias": "model-00001-of-00002.safetensors",
|
390 |
+
"gpt_neox.layers.9.mlp.dense_4h_to_h.weight": "model-00001-of-00002.safetensors",
|
391 |
+
"gpt_neox.layers.9.mlp.dense_h_to_4h.bias": "model-00001-of-00002.safetensors",
|
392 |
+
"gpt_neox.layers.9.mlp.dense_h_to_4h.weight": "model-00001-of-00002.safetensors",
|
393 |
+
"gpt_neox.layers.9.post_attention_layernorm.bias": "model-00001-of-00002.safetensors",
|
394 |
+
"gpt_neox.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors"
|
395 |
+
}
|
396 |
+
}
|
checkpoint-2000/rng_state_0.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:46345ffb3410dc8a643411e5653e0330e4e39d7b22bb179c8a35496aa6b2209d
|
3 |
+
size 15152
|
checkpoint-2000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cbb472e1bf798d638b02f0c0009cdfa99ce26db583db05edcfe91b9b86f5f1e7
|
3 |
+
size 1064
|
checkpoint-2000/special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<|endoftext|>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "<|endoftext|>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "<|endoftext|>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<|endoftext|>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
checkpoint-2000/tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
checkpoint-2000/tokenizer_config.json
ADDED
@@ -0,0 +1,214 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": false,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"add_prefix_space": false,
|
5 |
+
"added_tokens_decoder": {
|
6 |
+
"0": {
|
7 |
+
"content": "<|endoftext|>",
|
8 |
+
"lstrip": false,
|
9 |
+
"normalized": false,
|
10 |
+
"rstrip": false,
|
11 |
+
"single_word": false,
|
12 |
+
"special": true
|
13 |
+
},
|
14 |
+
"1": {
|
15 |
+
"content": "<|padding|>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": false,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false,
|
20 |
+
"special": true
|
21 |
+
},
|
22 |
+
"50254": {
|
23 |
+
"content": " ",
|
24 |
+
"lstrip": false,
|
25 |
+
"normalized": true,
|
26 |
+
"rstrip": false,
|
27 |
+
"single_word": false,
|
28 |
+
"special": false
|
29 |
+
},
|
30 |
+
"50255": {
|
31 |
+
"content": " ",
|
32 |
+
"lstrip": false,
|
33 |
+
"normalized": true,
|
34 |
+
"rstrip": false,
|
35 |
+
"single_word": false,
|
36 |
+
"special": false
|
37 |
+
},
|
38 |
+
"50256": {
|
39 |
+
"content": " ",
|
40 |
+
"lstrip": false,
|
41 |
+
"normalized": true,
|
42 |
+
"rstrip": false,
|
43 |
+
"single_word": false,
|
44 |
+
"special": false
|
45 |
+
},
|
46 |
+
"50257": {
|
47 |
+
"content": " ",
|
48 |
+
"lstrip": false,
|
49 |
+
"normalized": true,
|
50 |
+
"rstrip": false,
|
51 |
+
"single_word": false,
|
52 |
+
"special": false
|
53 |
+
},
|
54 |
+
"50258": {
|
55 |
+
"content": " ",
|
56 |
+
"lstrip": false,
|
57 |
+
"normalized": true,
|
58 |
+
"rstrip": false,
|
59 |
+
"single_word": false,
|
60 |
+
"special": false
|
61 |
+
},
|
62 |
+
"50259": {
|
63 |
+
"content": " ",
|
64 |
+
"lstrip": false,
|
65 |
+
"normalized": true,
|
66 |
+
"rstrip": false,
|
67 |
+
"single_word": false,
|
68 |
+
"special": false
|
69 |
+
},
|
70 |
+
"50260": {
|
71 |
+
"content": " ",
|
72 |
+
"lstrip": false,
|
73 |
+
"normalized": true,
|
74 |
+
"rstrip": false,
|
75 |
+
"single_word": false,
|
76 |
+
"special": false
|
77 |
+
},
|
78 |
+
"50261": {
|
79 |
+
"content": " ",
|
80 |
+
"lstrip": false,
|
81 |
+
"normalized": true,
|
82 |
+
"rstrip": false,
|
83 |
+
"single_word": false,
|
84 |
+
"special": false
|
85 |
+
},
|
86 |
+
"50262": {
|
87 |
+
"content": " ",
|
88 |
+
"lstrip": false,
|
89 |
+
"normalized": true,
|
90 |
+
"rstrip": false,
|
91 |
+
"single_word": false,
|
92 |
+
"special": false
|
93 |
+
},
|
94 |
+
"50263": {
|
95 |
+
"content": " ",
|
96 |
+
"lstrip": false,
|
97 |
+
"normalized": true,
|
98 |
+
"rstrip": false,
|
99 |
+
"single_word": false,
|
100 |
+
"special": false
|
101 |
+
},
|
102 |
+
"50264": {
|
103 |
+
"content": " ",
|
104 |
+
"lstrip": false,
|
105 |
+
"normalized": true,
|
106 |
+
"rstrip": false,
|
107 |
+
"single_word": false,
|
108 |
+
"special": false
|
109 |
+
},
|
110 |
+
"50265": {
|
111 |
+
"content": " ",
|
112 |
+
"lstrip": false,
|
113 |
+
"normalized": true,
|
114 |
+
"rstrip": false,
|
115 |
+
"single_word": false,
|
116 |
+
"special": false
|
117 |
+
},
|
118 |
+
"50266": {
|
119 |
+
"content": " ",
|
120 |
+
"lstrip": false,
|
121 |
+
"normalized": true,
|
122 |
+
"rstrip": false,
|
123 |
+
"single_word": false,
|
124 |
+
"special": false
|
125 |
+
},
|
126 |
+
"50267": {
|
127 |
+
"content": " ",
|
128 |
+
"lstrip": false,
|
129 |
+
"normalized": true,
|
130 |
+
"rstrip": false,
|
131 |
+
"single_word": false,
|
132 |
+
"special": false
|
133 |
+
},
|
134 |
+
"50268": {
|
135 |
+
"content": " ",
|
136 |
+
"lstrip": false,
|
137 |
+
"normalized": true,
|
138 |
+
"rstrip": false,
|
139 |
+
"single_word": false,
|
140 |
+
"special": false
|
141 |
+
},
|
142 |
+
"50269": {
|
143 |
+
"content": " ",
|
144 |
+
"lstrip": false,
|
145 |
+
"normalized": true,
|
146 |
+
"rstrip": false,
|
147 |
+
"single_word": false,
|
148 |
+
"special": false
|
149 |
+
},
|
150 |
+
"50270": {
|
151 |
+
"content": " ",
|
152 |
+
"lstrip": false,
|
153 |
+
"normalized": true,
|
154 |
+
"rstrip": false,
|
155 |
+
"single_word": false,
|
156 |
+
"special": false
|
157 |
+
},
|
158 |
+
"50271": {
|
159 |
+
"content": " ",
|
160 |
+
"lstrip": false,
|
161 |
+
"normalized": true,
|
162 |
+
"rstrip": false,
|
163 |
+
"single_word": false,
|
164 |
+
"special": false
|
165 |
+
},
|
166 |
+
"50272": {
|
167 |
+
"content": " ",
|
168 |
+
"lstrip": false,
|
169 |
+
"normalized": true,
|
170 |
+
"rstrip": false,
|
171 |
+
"single_word": false,
|
172 |
+
"special": false
|
173 |
+
},
|
174 |
+
"50273": {
|
175 |
+
"content": " ",
|
176 |
+
"lstrip": false,
|
177 |
+
"normalized": true,
|
178 |
+
"rstrip": false,
|
179 |
+
"single_word": false,
|
180 |
+
"special": false
|
181 |
+
},
|
182 |
+
"50274": {
|
183 |
+
"content": " ",
|
184 |
+
"lstrip": false,
|
185 |
+
"normalized": true,
|
186 |
+
"rstrip": false,
|
187 |
+
"single_word": false,
|
188 |
+
"special": false
|
189 |
+
},
|
190 |
+
"50275": {
|
191 |
+
"content": " ",
|
192 |
+
"lstrip": false,
|
193 |
+
"normalized": true,
|
194 |
+
"rstrip": false,
|
195 |
+
"single_word": false,
|
196 |
+
"special": false
|
197 |
+
},
|
198 |
+
"50276": {
|
199 |
+
"content": " ",
|
200 |
+
"lstrip": false,
|
201 |
+
"normalized": true,
|
202 |
+
"rstrip": false,
|
203 |
+
"single_word": false,
|
204 |
+
"special": false
|
205 |
+
}
|
206 |
+
},
|
207 |
+
"bos_token": "<|endoftext|>",
|
208 |
+
"clean_up_tokenization_spaces": true,
|
209 |
+
"eos_token": "<|endoftext|>",
|
210 |
+
"model_max_length": 1000000000000000019884624838656,
|
211 |
+
"pad_token": "<|endoftext|>",
|
212 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
213 |
+
"unk_token": "<|endoftext|>"
|
214 |
+
}
|
checkpoint-2000/trainer_state.json
ADDED
@@ -0,0 +1,1485 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 42.780748663101605,
|
5 |
+
"eval_steps": 400,
|
6 |
+
"global_step": 2000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.0213903743315508,
|
13 |
+
"grad_norm": 0.0,
|
14 |
+
"learning_rate": 0.0,
|
15 |
+
"loss": 1.8558,
|
16 |
+
"step": 1
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21390374331550802,
|
20 |
+
"grad_norm": 0.0,
|
21 |
+
"learning_rate": 0.0,
|
22 |
+
"loss": 2.0957,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.42780748663101603,
|
27 |
+
"grad_norm": 63.96761703491211,
|
28 |
+
"learning_rate": 8e-08,
|
29 |
+
"loss": 2.1274,
|
30 |
+
"step": 20
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.6417112299465241,
|
34 |
+
"grad_norm": 56.0494270324707,
|
35 |
+
"learning_rate": 4.4e-07,
|
36 |
+
"loss": 2.0513,
|
37 |
+
"step": 30
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.8556149732620321,
|
41 |
+
"grad_norm": 31.82117462158203,
|
42 |
+
"learning_rate": 8.400000000000001e-07,
|
43 |
+
"loss": 1.6275,
|
44 |
+
"step": 40
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 1.0695187165775402,
|
48 |
+
"grad_norm": 27.546764373779297,
|
49 |
+
"learning_rate": 1.2400000000000002e-06,
|
50 |
+
"loss": 1.3922,
|
51 |
+
"step": 50
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 1.2834224598930482,
|
55 |
+
"grad_norm": 27.145185470581055,
|
56 |
+
"learning_rate": 1.6400000000000002e-06,
|
57 |
+
"loss": 1.0424,
|
58 |
+
"step": 60
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 1.4973262032085561,
|
62 |
+
"grad_norm": 33.02227020263672,
|
63 |
+
"learning_rate": 2.04e-06,
|
64 |
+
"loss": 1.0434,
|
65 |
+
"step": 70
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 1.7112299465240641,
|
69 |
+
"grad_norm": 23.329740524291992,
|
70 |
+
"learning_rate": 2.4400000000000004e-06,
|
71 |
+
"loss": 1.0585,
|
72 |
+
"step": 80
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.9251336898395723,
|
76 |
+
"grad_norm": 26.408071517944336,
|
77 |
+
"learning_rate": 2.84e-06,
|
78 |
+
"loss": 0.8849,
|
79 |
+
"step": 90
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 2.1390374331550803,
|
83 |
+
"grad_norm": 17.790225982666016,
|
84 |
+
"learning_rate": 3.2400000000000003e-06,
|
85 |
+
"loss": 0.5427,
|
86 |
+
"step": 100
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 2.3529411764705883,
|
90 |
+
"grad_norm": 29.846942901611328,
|
91 |
+
"learning_rate": 3.6400000000000003e-06,
|
92 |
+
"loss": 0.2784,
|
93 |
+
"step": 110
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 2.5668449197860963,
|
97 |
+
"grad_norm": 20.77630043029785,
|
98 |
+
"learning_rate": 4.04e-06,
|
99 |
+
"loss": 0.2569,
|
100 |
+
"step": 120
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 2.7807486631016043,
|
104 |
+
"grad_norm": 33.25271987915039,
|
105 |
+
"learning_rate": 4.440000000000001e-06,
|
106 |
+
"loss": 0.1836,
|
107 |
+
"step": 130
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 2.9946524064171123,
|
111 |
+
"grad_norm": 66.23503875732422,
|
112 |
+
"learning_rate": 4.84e-06,
|
113 |
+
"loss": 0.3042,
|
114 |
+
"step": 140
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 3.2085561497326203,
|
118 |
+
"grad_norm": 4.618898868560791,
|
119 |
+
"learning_rate": 5.240000000000001e-06,
|
120 |
+
"loss": 0.0455,
|
121 |
+
"step": 150
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 3.4224598930481283,
|
125 |
+
"grad_norm": 5.856906890869141,
|
126 |
+
"learning_rate": 5.64e-06,
|
127 |
+
"loss": 0.0558,
|
128 |
+
"step": 160
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 3.6363636363636362,
|
132 |
+
"grad_norm": 25.529407501220703,
|
133 |
+
"learning_rate": 6.040000000000001e-06,
|
134 |
+
"loss": 0.0835,
|
135 |
+
"step": 170
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 3.8502673796791442,
|
139 |
+
"grad_norm": 16.971477508544922,
|
140 |
+
"learning_rate": 6.440000000000001e-06,
|
141 |
+
"loss": 0.1126,
|
142 |
+
"step": 180
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 4.064171122994653,
|
146 |
+
"grad_norm": 9.28547191619873,
|
147 |
+
"learning_rate": 6.8400000000000014e-06,
|
148 |
+
"loss": 0.0838,
|
149 |
+
"step": 190
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 4.278074866310161,
|
153 |
+
"grad_norm": 17.296798706054688,
|
154 |
+
"learning_rate": 7.24e-06,
|
155 |
+
"loss": 0.0518,
|
156 |
+
"step": 200
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 4.491978609625669,
|
160 |
+
"grad_norm": 12.538177490234375,
|
161 |
+
"learning_rate": 7.640000000000001e-06,
|
162 |
+
"loss": 0.0532,
|
163 |
+
"step": 210
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 4.705882352941177,
|
167 |
+
"grad_norm": 24.446659088134766,
|
168 |
+
"learning_rate": 8.040000000000001e-06,
|
169 |
+
"loss": 0.1583,
|
170 |
+
"step": 220
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 4.919786096256685,
|
174 |
+
"grad_norm": 25.139781951904297,
|
175 |
+
"learning_rate": 8.44e-06,
|
176 |
+
"loss": 0.0894,
|
177 |
+
"step": 230
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 5.133689839572193,
|
181 |
+
"grad_norm": 2.361607074737549,
|
182 |
+
"learning_rate": 8.84e-06,
|
183 |
+
"loss": 0.0531,
|
184 |
+
"step": 240
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 5.347593582887701,
|
188 |
+
"grad_norm": 4.768627643585205,
|
189 |
+
"learning_rate": 9.240000000000001e-06,
|
190 |
+
"loss": 0.0664,
|
191 |
+
"step": 250
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 5.561497326203209,
|
195 |
+
"grad_norm": 37.02785873413086,
|
196 |
+
"learning_rate": 9.640000000000001e-06,
|
197 |
+
"loss": 0.1502,
|
198 |
+
"step": 260
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 5.775401069518717,
|
202 |
+
"grad_norm": 48.693763732910156,
|
203 |
+
"learning_rate": 1e-05,
|
204 |
+
"loss": 0.1953,
|
205 |
+
"step": 270
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 5.989304812834225,
|
209 |
+
"grad_norm": 37.84030532836914,
|
210 |
+
"learning_rate": 1e-05,
|
211 |
+
"loss": 0.1806,
|
212 |
+
"step": 280
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 6.2032085561497325,
|
216 |
+
"grad_norm": 39.74201965332031,
|
217 |
+
"learning_rate": 1e-05,
|
218 |
+
"loss": 0.0866,
|
219 |
+
"step": 290
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 6.4171122994652405,
|
223 |
+
"grad_norm": 7.8511834144592285,
|
224 |
+
"learning_rate": 1e-05,
|
225 |
+
"loss": 0.0994,
|
226 |
+
"step": 300
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 6.6310160427807485,
|
230 |
+
"grad_norm": 54.63271713256836,
|
231 |
+
"learning_rate": 1e-05,
|
232 |
+
"loss": 0.1712,
|
233 |
+
"step": 310
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 6.8449197860962565,
|
237 |
+
"grad_norm": 44.596153259277344,
|
238 |
+
"learning_rate": 1e-05,
|
239 |
+
"loss": 0.1442,
|
240 |
+
"step": 320
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 7.0588235294117645,
|
244 |
+
"grad_norm": 5.094964504241943,
|
245 |
+
"learning_rate": 1e-05,
|
246 |
+
"loss": 0.0526,
|
247 |
+
"step": 330
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 7.2727272727272725,
|
251 |
+
"grad_norm": 12.109665870666504,
|
252 |
+
"learning_rate": 1e-05,
|
253 |
+
"loss": 0.03,
|
254 |
+
"step": 340
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 7.4866310160427805,
|
258 |
+
"grad_norm": 34.3682975769043,
|
259 |
+
"learning_rate": 1e-05,
|
260 |
+
"loss": 0.0759,
|
261 |
+
"step": 350
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 7.7005347593582885,
|
265 |
+
"grad_norm": 2.3397929668426514,
|
266 |
+
"learning_rate": 1e-05,
|
267 |
+
"loss": 0.0685,
|
268 |
+
"step": 360
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 7.9144385026737964,
|
272 |
+
"grad_norm": 21.006542205810547,
|
273 |
+
"learning_rate": 1e-05,
|
274 |
+
"loss": 0.0637,
|
275 |
+
"step": 370
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 8.128342245989305,
|
279 |
+
"grad_norm": 2.5418567657470703,
|
280 |
+
"learning_rate": 1e-05,
|
281 |
+
"loss": 0.057,
|
282 |
+
"step": 380
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 8.342245989304812,
|
286 |
+
"grad_norm": 18.59604263305664,
|
287 |
+
"learning_rate": 1e-05,
|
288 |
+
"loss": 0.1096,
|
289 |
+
"step": 390
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 8.556149732620321,
|
293 |
+
"grad_norm": 52.85743713378906,
|
294 |
+
"learning_rate": 1e-05,
|
295 |
+
"loss": 0.1422,
|
296 |
+
"step": 400
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 8.556149732620321,
|
300 |
+
"eval_accuracy": 0.4,
|
301 |
+
"eval_loss": 4.6875,
|
302 |
+
"eval_runtime": 0.8724,
|
303 |
+
"eval_samples_per_second": 11.462,
|
304 |
+
"eval_steps_per_second": 1.146,
|
305 |
+
"step": 400
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 8.770053475935828,
|
309 |
+
"grad_norm": 36.897674560546875,
|
310 |
+
"learning_rate": 1e-05,
|
311 |
+
"loss": 0.1633,
|
312 |
+
"step": 410
|
313 |
+
},
|
314 |
+
{
|
315 |
+
"epoch": 8.983957219251337,
|
316 |
+
"grad_norm": 16.528366088867188,
|
317 |
+
"learning_rate": 1e-05,
|
318 |
+
"loss": 0.0336,
|
319 |
+
"step": 420
|
320 |
+
},
|
321 |
+
{
|
322 |
+
"epoch": 9.197860962566844,
|
323 |
+
"grad_norm": 87.09658813476562,
|
324 |
+
"learning_rate": 1e-05,
|
325 |
+
"loss": 0.1071,
|
326 |
+
"step": 430
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 9.411764705882353,
|
330 |
+
"grad_norm": 4.741786479949951,
|
331 |
+
"learning_rate": 1e-05,
|
332 |
+
"loss": 0.0701,
|
333 |
+
"step": 440
|
334 |
+
},
|
335 |
+
{
|
336 |
+
"epoch": 9.62566844919786,
|
337 |
+
"grad_norm": 12.09588623046875,
|
338 |
+
"learning_rate": 1e-05,
|
339 |
+
"loss": 0.0746,
|
340 |
+
"step": 450
|
341 |
+
},
|
342 |
+
{
|
343 |
+
"epoch": 9.83957219251337,
|
344 |
+
"grad_norm": 32.604068756103516,
|
345 |
+
"learning_rate": 1e-05,
|
346 |
+
"loss": 0.0872,
|
347 |
+
"step": 460
|
348 |
+
},
|
349 |
+
{
|
350 |
+
"epoch": 10.053475935828876,
|
351 |
+
"grad_norm": 4.408775806427002,
|
352 |
+
"learning_rate": 1e-05,
|
353 |
+
"loss": 0.1084,
|
354 |
+
"step": 470
|
355 |
+
},
|
356 |
+
{
|
357 |
+
"epoch": 10.267379679144385,
|
358 |
+
"grad_norm": 22.274890899658203,
|
359 |
+
"learning_rate": 1e-05,
|
360 |
+
"loss": 0.0297,
|
361 |
+
"step": 480
|
362 |
+
},
|
363 |
+
{
|
364 |
+
"epoch": 10.481283422459892,
|
365 |
+
"grad_norm": 58.223323822021484,
|
366 |
+
"learning_rate": 1e-05,
|
367 |
+
"loss": 0.0637,
|
368 |
+
"step": 490
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 10.695187165775401,
|
372 |
+
"grad_norm": 2.7027640342712402,
|
373 |
+
"learning_rate": 1e-05,
|
374 |
+
"loss": 0.1206,
|
375 |
+
"step": 500
|
376 |
+
},
|
377 |
+
{
|
378 |
+
"epoch": 10.909090909090908,
|
379 |
+
"grad_norm": 43.47382736206055,
|
380 |
+
"learning_rate": 1e-05,
|
381 |
+
"loss": 0.045,
|
382 |
+
"step": 510
|
383 |
+
},
|
384 |
+
{
|
385 |
+
"epoch": 11.122994652406417,
|
386 |
+
"grad_norm": 1.6665784120559692,
|
387 |
+
"learning_rate": 1e-05,
|
388 |
+
"loss": 0.045,
|
389 |
+
"step": 520
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 11.336898395721924,
|
393 |
+
"grad_norm": 0.5738474130630493,
|
394 |
+
"learning_rate": 1e-05,
|
395 |
+
"loss": 0.008,
|
396 |
+
"step": 530
|
397 |
+
},
|
398 |
+
{
|
399 |
+
"epoch": 11.550802139037433,
|
400 |
+
"grad_norm": 16.410573959350586,
|
401 |
+
"learning_rate": 1e-05,
|
402 |
+
"loss": 0.0995,
|
403 |
+
"step": 540
|
404 |
+
},
|
405 |
+
{
|
406 |
+
"epoch": 11.764705882352942,
|
407 |
+
"grad_norm": 0.7015154361724854,
|
408 |
+
"learning_rate": 1e-05,
|
409 |
+
"loss": 0.0313,
|
410 |
+
"step": 550
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 11.97860962566845,
|
414 |
+
"grad_norm": 11.851508140563965,
|
415 |
+
"learning_rate": 1e-05,
|
416 |
+
"loss": 0.0724,
|
417 |
+
"step": 560
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 12.192513368983958,
|
421 |
+
"grad_norm": 16.754575729370117,
|
422 |
+
"learning_rate": 1e-05,
|
423 |
+
"loss": 0.0272,
|
424 |
+
"step": 570
|
425 |
+
},
|
426 |
+
{
|
427 |
+
"epoch": 12.406417112299465,
|
428 |
+
"grad_norm": 15.92441177368164,
|
429 |
+
"learning_rate": 1e-05,
|
430 |
+
"loss": 0.0241,
|
431 |
+
"step": 580
|
432 |
+
},
|
433 |
+
{
|
434 |
+
"epoch": 12.620320855614974,
|
435 |
+
"grad_norm": 0.05043609067797661,
|
436 |
+
"learning_rate": 1e-05,
|
437 |
+
"loss": 0.042,
|
438 |
+
"step": 590
|
439 |
+
},
|
440 |
+
{
|
441 |
+
"epoch": 12.834224598930481,
|
442 |
+
"grad_norm": 0.3494648337364197,
|
443 |
+
"learning_rate": 1e-05,
|
444 |
+
"loss": 0.0083,
|
445 |
+
"step": 600
|
446 |
+
},
|
447 |
+
{
|
448 |
+
"epoch": 13.04812834224599,
|
449 |
+
"grad_norm": 23.260435104370117,
|
450 |
+
"learning_rate": 1e-05,
|
451 |
+
"loss": 0.026,
|
452 |
+
"step": 610
|
453 |
+
},
|
454 |
+
{
|
455 |
+
"epoch": 13.262032085561497,
|
456 |
+
"grad_norm": 5.789357662200928,
|
457 |
+
"learning_rate": 1e-05,
|
458 |
+
"loss": 0.0224,
|
459 |
+
"step": 620
|
460 |
+
},
|
461 |
+
{
|
462 |
+
"epoch": 13.475935828877006,
|
463 |
+
"grad_norm": 0.056954510509967804,
|
464 |
+
"learning_rate": 1e-05,
|
465 |
+
"loss": 0.0642,
|
466 |
+
"step": 630
|
467 |
+
},
|
468 |
+
{
|
469 |
+
"epoch": 13.689839572192513,
|
470 |
+
"grad_norm": 72.28430938720703,
|
471 |
+
"learning_rate": 1e-05,
|
472 |
+
"loss": 0.0366,
|
473 |
+
"step": 640
|
474 |
+
},
|
475 |
+
{
|
476 |
+
"epoch": 13.903743315508022,
|
477 |
+
"grad_norm": 0.5189899802207947,
|
478 |
+
"learning_rate": 1e-05,
|
479 |
+
"loss": 0.033,
|
480 |
+
"step": 650
|
481 |
+
},
|
482 |
+
{
|
483 |
+
"epoch": 14.117647058823529,
|
484 |
+
"grad_norm": 0.07448354363441467,
|
485 |
+
"learning_rate": 1e-05,
|
486 |
+
"loss": 0.0088,
|
487 |
+
"step": 660
|
488 |
+
},
|
489 |
+
{
|
490 |
+
"epoch": 14.331550802139038,
|
491 |
+
"grad_norm": 0.05018957331776619,
|
492 |
+
"learning_rate": 1e-05,
|
493 |
+
"loss": 0.0082,
|
494 |
+
"step": 670
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 14.545454545454545,
|
498 |
+
"grad_norm": 33.61125183105469,
|
499 |
+
"learning_rate": 1e-05,
|
500 |
+
"loss": 0.0753,
|
501 |
+
"step": 680
|
502 |
+
},
|
503 |
+
{
|
504 |
+
"epoch": 14.759358288770054,
|
505 |
+
"grad_norm": 0.023922119289636612,
|
506 |
+
"learning_rate": 1e-05,
|
507 |
+
"loss": 0.0267,
|
508 |
+
"step": 690
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 14.973262032085561,
|
512 |
+
"grad_norm": 0.04713653773069382,
|
513 |
+
"learning_rate": 1e-05,
|
514 |
+
"loss": 0.096,
|
515 |
+
"step": 700
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 15.18716577540107,
|
519 |
+
"grad_norm": 0.0457034632563591,
|
520 |
+
"learning_rate": 1e-05,
|
521 |
+
"loss": 0.0083,
|
522 |
+
"step": 710
|
523 |
+
},
|
524 |
+
{
|
525 |
+
"epoch": 15.401069518716577,
|
526 |
+
"grad_norm": 0.011708029545843601,
|
527 |
+
"learning_rate": 1e-05,
|
528 |
+
"loss": 0.072,
|
529 |
+
"step": 720
|
530 |
+
},
|
531 |
+
{
|
532 |
+
"epoch": 15.614973262032086,
|
533 |
+
"grad_norm": 0.3412806987762451,
|
534 |
+
"learning_rate": 1e-05,
|
535 |
+
"loss": 0.0032,
|
536 |
+
"step": 730
|
537 |
+
},
|
538 |
+
{
|
539 |
+
"epoch": 15.828877005347593,
|
540 |
+
"grad_norm": 1.667181372642517,
|
541 |
+
"learning_rate": 1e-05,
|
542 |
+
"loss": 0.0512,
|
543 |
+
"step": 740
|
544 |
+
},
|
545 |
+
{
|
546 |
+
"epoch": 16.0427807486631,
|
547 |
+
"grad_norm": 0.004597651772201061,
|
548 |
+
"learning_rate": 1e-05,
|
549 |
+
"loss": 0.0555,
|
550 |
+
"step": 750
|
551 |
+
},
|
552 |
+
{
|
553 |
+
"epoch": 16.25668449197861,
|
554 |
+
"grad_norm": 0.01679002121090889,
|
555 |
+
"learning_rate": 1e-05,
|
556 |
+
"loss": 0.0294,
|
557 |
+
"step": 760
|
558 |
+
},
|
559 |
+
{
|
560 |
+
"epoch": 16.470588235294116,
|
561 |
+
"grad_norm": 24.51487922668457,
|
562 |
+
"learning_rate": 1e-05,
|
563 |
+
"loss": 0.0979,
|
564 |
+
"step": 770
|
565 |
+
},
|
566 |
+
{
|
567 |
+
"epoch": 16.684491978609625,
|
568 |
+
"grad_norm": 0.010406852699816227,
|
569 |
+
"learning_rate": 1e-05,
|
570 |
+
"loss": 0.0361,
|
571 |
+
"step": 780
|
572 |
+
},
|
573 |
+
{
|
574 |
+
"epoch": 16.898395721925134,
|
575 |
+
"grad_norm": 4.567187786102295,
|
576 |
+
"learning_rate": 1e-05,
|
577 |
+
"loss": 0.1229,
|
578 |
+
"step": 790
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 17.112299465240643,
|
582 |
+
"grad_norm": 1.7842605113983154,
|
583 |
+
"learning_rate": 1e-05,
|
584 |
+
"loss": 0.0121,
|
585 |
+
"step": 800
|
586 |
+
},
|
587 |
+
{
|
588 |
+
"epoch": 17.112299465240643,
|
589 |
+
"eval_accuracy": 0.5,
|
590 |
+
"eval_loss": 6.6640625,
|
591 |
+
"eval_runtime": 0.8689,
|
592 |
+
"eval_samples_per_second": 11.509,
|
593 |
+
"eval_steps_per_second": 1.151,
|
594 |
+
"step": 800
|
595 |
+
},
|
596 |
+
{
|
597 |
+
"epoch": 17.32620320855615,
|
598 |
+
"grad_norm": 0.061683282256126404,
|
599 |
+
"learning_rate": 1e-05,
|
600 |
+
"loss": 0.0079,
|
601 |
+
"step": 810
|
602 |
+
},
|
603 |
+
{
|
604 |
+
"epoch": 17.540106951871657,
|
605 |
+
"grad_norm": 0.11749367415904999,
|
606 |
+
"learning_rate": 1e-05,
|
607 |
+
"loss": 0.0857,
|
608 |
+
"step": 820
|
609 |
+
},
|
610 |
+
{
|
611 |
+
"epoch": 17.754010695187166,
|
612 |
+
"grad_norm": 0.02787230722606182,
|
613 |
+
"learning_rate": 1e-05,
|
614 |
+
"loss": 0.0051,
|
615 |
+
"step": 830
|
616 |
+
},
|
617 |
+
{
|
618 |
+
"epoch": 17.967914438502675,
|
619 |
+
"grad_norm": 0.015304004773497581,
|
620 |
+
"learning_rate": 1e-05,
|
621 |
+
"loss": 0.0882,
|
622 |
+
"step": 840
|
623 |
+
},
|
624 |
+
{
|
625 |
+
"epoch": 18.181818181818183,
|
626 |
+
"grad_norm": 0.41026026010513306,
|
627 |
+
"learning_rate": 1e-05,
|
628 |
+
"loss": 0.0002,
|
629 |
+
"step": 850
|
630 |
+
},
|
631 |
+
{
|
632 |
+
"epoch": 18.39572192513369,
|
633 |
+
"grad_norm": 0.012662127614021301,
|
634 |
+
"learning_rate": 1e-05,
|
635 |
+
"loss": 0.0013,
|
636 |
+
"step": 860
|
637 |
+
},
|
638 |
+
{
|
639 |
+
"epoch": 18.609625668449198,
|
640 |
+
"grad_norm": 0.10844732820987701,
|
641 |
+
"learning_rate": 1e-05,
|
642 |
+
"loss": 0.0062,
|
643 |
+
"step": 870
|
644 |
+
},
|
645 |
+
{
|
646 |
+
"epoch": 18.823529411764707,
|
647 |
+
"grad_norm": 2.6394922733306885,
|
648 |
+
"learning_rate": 1e-05,
|
649 |
+
"loss": 0.0335,
|
650 |
+
"step": 880
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 19.037433155080215,
|
654 |
+
"grad_norm": 0.28504815697669983,
|
655 |
+
"learning_rate": 1e-05,
|
656 |
+
"loss": 0.0092,
|
657 |
+
"step": 890
|
658 |
+
},
|
659 |
+
{
|
660 |
+
"epoch": 19.25133689839572,
|
661 |
+
"grad_norm": 0.03371066227555275,
|
662 |
+
"learning_rate": 1e-05,
|
663 |
+
"loss": 0.0001,
|
664 |
+
"step": 900
|
665 |
+
},
|
666 |
+
{
|
667 |
+
"epoch": 19.46524064171123,
|
668 |
+
"grad_norm": 0.01855628192424774,
|
669 |
+
"learning_rate": 1e-05,
|
670 |
+
"loss": 0.0046,
|
671 |
+
"step": 910
|
672 |
+
},
|
673 |
+
{
|
674 |
+
"epoch": 19.67914438502674,
|
675 |
+
"grad_norm": 0.008938683196902275,
|
676 |
+
"learning_rate": 1e-05,
|
677 |
+
"loss": 0.0014,
|
678 |
+
"step": 920
|
679 |
+
},
|
680 |
+
{
|
681 |
+
"epoch": 19.893048128342247,
|
682 |
+
"grad_norm": 0.0025397331919521093,
|
683 |
+
"learning_rate": 1e-05,
|
684 |
+
"loss": 0.0023,
|
685 |
+
"step": 930
|
686 |
+
},
|
687 |
+
{
|
688 |
+
"epoch": 20.106951871657753,
|
689 |
+
"grad_norm": 0.0024771266616880894,
|
690 |
+
"learning_rate": 1e-05,
|
691 |
+
"loss": 0.001,
|
692 |
+
"step": 940
|
693 |
+
},
|
694 |
+
{
|
695 |
+
"epoch": 20.32085561497326,
|
696 |
+
"grad_norm": 56.46254348754883,
|
697 |
+
"learning_rate": 1e-05,
|
698 |
+
"loss": 0.0533,
|
699 |
+
"step": 950
|
700 |
+
},
|
701 |
+
{
|
702 |
+
"epoch": 20.53475935828877,
|
703 |
+
"grad_norm": 0.046234406530857086,
|
704 |
+
"learning_rate": 1e-05,
|
705 |
+
"loss": 0.0113,
|
706 |
+
"step": 960
|
707 |
+
},
|
708 |
+
{
|
709 |
+
"epoch": 20.74866310160428,
|
710 |
+
"grad_norm": 0.03888735547661781,
|
711 |
+
"learning_rate": 1e-05,
|
712 |
+
"loss": 0.0001,
|
713 |
+
"step": 970
|
714 |
+
},
|
715 |
+
{
|
716 |
+
"epoch": 20.962566844919785,
|
717 |
+
"grad_norm": 4.142932415008545,
|
718 |
+
"learning_rate": 1e-05,
|
719 |
+
"loss": 0.0078,
|
720 |
+
"step": 980
|
721 |
+
},
|
722 |
+
{
|
723 |
+
"epoch": 21.176470588235293,
|
724 |
+
"grad_norm": 0.01926487125456333,
|
725 |
+
"learning_rate": 1e-05,
|
726 |
+
"loss": 0.0015,
|
727 |
+
"step": 990
|
728 |
+
},
|
729 |
+
{
|
730 |
+
"epoch": 21.390374331550802,
|
731 |
+
"grad_norm": 0.03864801302552223,
|
732 |
+
"learning_rate": 1e-05,
|
733 |
+
"loss": 0.0032,
|
734 |
+
"step": 1000
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 21.60427807486631,
|
738 |
+
"grad_norm": 0.40588998794555664,
|
739 |
+
"learning_rate": 1e-05,
|
740 |
+
"loss": 0.0003,
|
741 |
+
"step": 1010
|
742 |
+
},
|
743 |
+
{
|
744 |
+
"epoch": 21.818181818181817,
|
745 |
+
"grad_norm": 0.2934487760066986,
|
746 |
+
"learning_rate": 1e-05,
|
747 |
+
"loss": 0.0081,
|
748 |
+
"step": 1020
|
749 |
+
},
|
750 |
+
{
|
751 |
+
"epoch": 22.032085561497325,
|
752 |
+
"grad_norm": 0.004641254432499409,
|
753 |
+
"learning_rate": 1e-05,
|
754 |
+
"loss": 0.0,
|
755 |
+
"step": 1030
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 22.245989304812834,
|
759 |
+
"grad_norm": 14.084956169128418,
|
760 |
+
"learning_rate": 1e-05,
|
761 |
+
"loss": 0.0018,
|
762 |
+
"step": 1040
|
763 |
+
},
|
764 |
+
{
|
765 |
+
"epoch": 22.459893048128343,
|
766 |
+
"grad_norm": 1.7572118043899536,
|
767 |
+
"learning_rate": 1e-05,
|
768 |
+
"loss": 0.001,
|
769 |
+
"step": 1050
|
770 |
+
},
|
771 |
+
{
|
772 |
+
"epoch": 22.67379679144385,
|
773 |
+
"grad_norm": 2.7290031909942627,
|
774 |
+
"learning_rate": 1e-05,
|
775 |
+
"loss": 0.0002,
|
776 |
+
"step": 1060
|
777 |
+
},
|
778 |
+
{
|
779 |
+
"epoch": 22.887700534759357,
|
780 |
+
"grad_norm": 0.005051525309681892,
|
781 |
+
"learning_rate": 1e-05,
|
782 |
+
"loss": 0.0007,
|
783 |
+
"step": 1070
|
784 |
+
},
|
785 |
+
{
|
786 |
+
"epoch": 23.101604278074866,
|
787 |
+
"grad_norm": 0.002047200920060277,
|
788 |
+
"learning_rate": 1e-05,
|
789 |
+
"loss": 0.0,
|
790 |
+
"step": 1080
|
791 |
+
},
|
792 |
+
{
|
793 |
+
"epoch": 23.315508021390375,
|
794 |
+
"grad_norm": 0.0035107170697301626,
|
795 |
+
"learning_rate": 1e-05,
|
796 |
+
"loss": 0.0494,
|
797 |
+
"step": 1090
|
798 |
+
},
|
799 |
+
{
|
800 |
+
"epoch": 23.529411764705884,
|
801 |
+
"grad_norm": 0.004472868517041206,
|
802 |
+
"learning_rate": 1e-05,
|
803 |
+
"loss": 0.0034,
|
804 |
+
"step": 1100
|
805 |
+
},
|
806 |
+
{
|
807 |
+
"epoch": 23.74331550802139,
|
808 |
+
"grad_norm": 0.08594793826341629,
|
809 |
+
"learning_rate": 1e-05,
|
810 |
+
"loss": 0.0063,
|
811 |
+
"step": 1110
|
812 |
+
},
|
813 |
+
{
|
814 |
+
"epoch": 23.9572192513369,
|
815 |
+
"grad_norm": 0.752574622631073,
|
816 |
+
"learning_rate": 1e-05,
|
817 |
+
"loss": 0.0002,
|
818 |
+
"step": 1120
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 24.171122994652407,
|
822 |
+
"grad_norm": 0.021934740245342255,
|
823 |
+
"learning_rate": 1e-05,
|
824 |
+
"loss": 0.0019,
|
825 |
+
"step": 1130
|
826 |
+
},
|
827 |
+
{
|
828 |
+
"epoch": 24.385026737967916,
|
829 |
+
"grad_norm": 0.007998108863830566,
|
830 |
+
"learning_rate": 1e-05,
|
831 |
+
"loss": 0.0,
|
832 |
+
"step": 1140
|
833 |
+
},
|
834 |
+
{
|
835 |
+
"epoch": 24.59893048128342,
|
836 |
+
"grad_norm": 0.010390130802989006,
|
837 |
+
"learning_rate": 1e-05,
|
838 |
+
"loss": 0.0072,
|
839 |
+
"step": 1150
|
840 |
+
},
|
841 |
+
{
|
842 |
+
"epoch": 24.81283422459893,
|
843 |
+
"grad_norm": 0.002071298658847809,
|
844 |
+
"learning_rate": 1e-05,
|
845 |
+
"loss": 0.0013,
|
846 |
+
"step": 1160
|
847 |
+
},
|
848 |
+
{
|
849 |
+
"epoch": 25.02673796791444,
|
850 |
+
"grad_norm": 0.0045625255443155766,
|
851 |
+
"learning_rate": 1e-05,
|
852 |
+
"loss": 0.002,
|
853 |
+
"step": 1170
|
854 |
+
},
|
855 |
+
{
|
856 |
+
"epoch": 25.240641711229948,
|
857 |
+
"grad_norm": 0.017212701961398125,
|
858 |
+
"learning_rate": 1e-05,
|
859 |
+
"loss": 0.0048,
|
860 |
+
"step": 1180
|
861 |
+
},
|
862 |
+
{
|
863 |
+
"epoch": 25.454545454545453,
|
864 |
+
"grad_norm": 12.734994888305664,
|
865 |
+
"learning_rate": 1e-05,
|
866 |
+
"loss": 0.0018,
|
867 |
+
"step": 1190
|
868 |
+
},
|
869 |
+
{
|
870 |
+
"epoch": 25.668449197860962,
|
871 |
+
"grad_norm": 0.0035849984269589186,
|
872 |
+
"learning_rate": 1e-05,
|
873 |
+
"loss": 0.0,
|
874 |
+
"step": 1200
|
875 |
+
},
|
876 |
+
{
|
877 |
+
"epoch": 25.668449197860962,
|
878 |
+
"eval_accuracy": 0.8,
|
879 |
+
"eval_loss": 4.71484375,
|
880 |
+
"eval_runtime": 0.8712,
|
881 |
+
"eval_samples_per_second": 11.478,
|
882 |
+
"eval_steps_per_second": 1.148,
|
883 |
+
"step": 1200
|
884 |
+
},
|
885 |
+
{
|
886 |
+
"epoch": 25.88235294117647,
|
887 |
+
"grad_norm": 0.005425190087407827,
|
888 |
+
"learning_rate": 1e-05,
|
889 |
+
"loss": 0.0,
|
890 |
+
"step": 1210
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 26.09625668449198,
|
894 |
+
"grad_norm": 0.007795536890625954,
|
895 |
+
"learning_rate": 1e-05,
|
896 |
+
"loss": 0.0034,
|
897 |
+
"step": 1220
|
898 |
+
},
|
899 |
+
{
|
900 |
+
"epoch": 26.310160427807485,
|
901 |
+
"grad_norm": 0.00858351681381464,
|
902 |
+
"learning_rate": 1e-05,
|
903 |
+
"loss": 0.0032,
|
904 |
+
"step": 1230
|
905 |
+
},
|
906 |
+
{
|
907 |
+
"epoch": 26.524064171122994,
|
908 |
+
"grad_norm": 0.0035222836304455996,
|
909 |
+
"learning_rate": 1e-05,
|
910 |
+
"loss": 0.0103,
|
911 |
+
"step": 1240
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 26.737967914438503,
|
915 |
+
"grad_norm": 0.004781092517077923,
|
916 |
+
"learning_rate": 1e-05,
|
917 |
+
"loss": 0.0002,
|
918 |
+
"step": 1250
|
919 |
+
},
|
920 |
+
{
|
921 |
+
"epoch": 26.951871657754012,
|
922 |
+
"grad_norm": 0.0031953530851751566,
|
923 |
+
"learning_rate": 1e-05,
|
924 |
+
"loss": 0.0013,
|
925 |
+
"step": 1260
|
926 |
+
},
|
927 |
+
{
|
928 |
+
"epoch": 27.165775401069517,
|
929 |
+
"grad_norm": 0.00516524026170373,
|
930 |
+
"learning_rate": 1e-05,
|
931 |
+
"loss": 0.0007,
|
932 |
+
"step": 1270
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 27.379679144385026,
|
936 |
+
"grad_norm": 0.004281027242541313,
|
937 |
+
"learning_rate": 1e-05,
|
938 |
+
"loss": 0.0106,
|
939 |
+
"step": 1280
|
940 |
+
},
|
941 |
+
{
|
942 |
+
"epoch": 27.593582887700535,
|
943 |
+
"grad_norm": 0.0015023744199424982,
|
944 |
+
"learning_rate": 1e-05,
|
945 |
+
"loss": 0.0,
|
946 |
+
"step": 1290
|
947 |
+
},
|
948 |
+
{
|
949 |
+
"epoch": 27.807486631016044,
|
950 |
+
"grad_norm": 0.005484332330524921,
|
951 |
+
"learning_rate": 1e-05,
|
952 |
+
"loss": 0.0023,
|
953 |
+
"step": 1300
|
954 |
+
},
|
955 |
+
{
|
956 |
+
"epoch": 28.02139037433155,
|
957 |
+
"grad_norm": 0.0035302124451845884,
|
958 |
+
"learning_rate": 1e-05,
|
959 |
+
"loss": 0.0001,
|
960 |
+
"step": 1310
|
961 |
+
},
|
962 |
+
{
|
963 |
+
"epoch": 28.235294117647058,
|
964 |
+
"grad_norm": 0.0033441612031310797,
|
965 |
+
"learning_rate": 1e-05,
|
966 |
+
"loss": 0.0025,
|
967 |
+
"step": 1320
|
968 |
+
},
|
969 |
+
{
|
970 |
+
"epoch": 28.449197860962567,
|
971 |
+
"grad_norm": 0.0035446221008896828,
|
972 |
+
"learning_rate": 1e-05,
|
973 |
+
"loss": 0.0014,
|
974 |
+
"step": 1330
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 28.663101604278076,
|
978 |
+
"grad_norm": 0.010898416861891747,
|
979 |
+
"learning_rate": 1e-05,
|
980 |
+
"loss": 0.002,
|
981 |
+
"step": 1340
|
982 |
+
},
|
983 |
+
{
|
984 |
+
"epoch": 28.87700534759358,
|
985 |
+
"grad_norm": 0.008727981708943844,
|
986 |
+
"learning_rate": 1e-05,
|
987 |
+
"loss": 0.1828,
|
988 |
+
"step": 1350
|
989 |
+
},
|
990 |
+
{
|
991 |
+
"epoch": 29.09090909090909,
|
992 |
+
"grad_norm": 0.004256559070199728,
|
993 |
+
"learning_rate": 1e-05,
|
994 |
+
"loss": 0.0431,
|
995 |
+
"step": 1360
|
996 |
+
},
|
997 |
+
{
|
998 |
+
"epoch": 29.3048128342246,
|
999 |
+
"grad_norm": 0.21178306639194489,
|
1000 |
+
"learning_rate": 1e-05,
|
1001 |
+
"loss": 0.0099,
|
1002 |
+
"step": 1370
|
1003 |
+
},
|
1004 |
+
{
|
1005 |
+
"epoch": 29.518716577540108,
|
1006 |
+
"grad_norm": 0.2352713793516159,
|
1007 |
+
"learning_rate": 1e-05,
|
1008 |
+
"loss": 0.0835,
|
1009 |
+
"step": 1380
|
1010 |
+
},
|
1011 |
+
{
|
1012 |
+
"epoch": 29.732620320855617,
|
1013 |
+
"grad_norm": 0.03791838884353638,
|
1014 |
+
"learning_rate": 1e-05,
|
1015 |
+
"loss": 0.0522,
|
1016 |
+
"step": 1390
|
1017 |
+
},
|
1018 |
+
{
|
1019 |
+
"epoch": 29.946524064171122,
|
1020 |
+
"grad_norm": 0.04343949630856514,
|
1021 |
+
"learning_rate": 1e-05,
|
1022 |
+
"loss": 0.0004,
|
1023 |
+
"step": 1400
|
1024 |
+
},
|
1025 |
+
{
|
1026 |
+
"epoch": 30.16042780748663,
|
1027 |
+
"grad_norm": 2.7350242137908936,
|
1028 |
+
"learning_rate": 1e-05,
|
1029 |
+
"loss": 0.003,
|
1030 |
+
"step": 1410
|
1031 |
+
},
|
1032 |
+
{
|
1033 |
+
"epoch": 30.37433155080214,
|
1034 |
+
"grad_norm": 0.030035167932510376,
|
1035 |
+
"learning_rate": 1e-05,
|
1036 |
+
"loss": 0.006,
|
1037 |
+
"step": 1420
|
1038 |
+
},
|
1039 |
+
{
|
1040 |
+
"epoch": 30.58823529411765,
|
1041 |
+
"grad_norm": 3.1150763034820557,
|
1042 |
+
"learning_rate": 1e-05,
|
1043 |
+
"loss": 0.0016,
|
1044 |
+
"step": 1430
|
1045 |
+
},
|
1046 |
+
{
|
1047 |
+
"epoch": 30.802139037433154,
|
1048 |
+
"grad_norm": 0.028651399537920952,
|
1049 |
+
"learning_rate": 1e-05,
|
1050 |
+
"loss": 0.0016,
|
1051 |
+
"step": 1440
|
1052 |
+
},
|
1053 |
+
{
|
1054 |
+
"epoch": 31.016042780748663,
|
1055 |
+
"grad_norm": 37.155609130859375,
|
1056 |
+
"learning_rate": 1e-05,
|
1057 |
+
"loss": 0.0128,
|
1058 |
+
"step": 1450
|
1059 |
+
},
|
1060 |
+
{
|
1061 |
+
"epoch": 31.22994652406417,
|
1062 |
+
"grad_norm": 0.002968219807371497,
|
1063 |
+
"learning_rate": 1e-05,
|
1064 |
+
"loss": 0.0141,
|
1065 |
+
"step": 1460
|
1066 |
+
},
|
1067 |
+
{
|
1068 |
+
"epoch": 31.44385026737968,
|
1069 |
+
"grad_norm": 5.24705696105957,
|
1070 |
+
"learning_rate": 1e-05,
|
1071 |
+
"loss": 0.0121,
|
1072 |
+
"step": 1470
|
1073 |
+
},
|
1074 |
+
{
|
1075 |
+
"epoch": 31.657754010695186,
|
1076 |
+
"grad_norm": 0.0024348751176148653,
|
1077 |
+
"learning_rate": 1e-05,
|
1078 |
+
"loss": 0.1083,
|
1079 |
+
"step": 1480
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 31.871657754010695,
|
1083 |
+
"grad_norm": 0.5254661440849304,
|
1084 |
+
"learning_rate": 1e-05,
|
1085 |
+
"loss": 0.1179,
|
1086 |
+
"step": 1490
|
1087 |
+
},
|
1088 |
+
{
|
1089 |
+
"epoch": 32.0855614973262,
|
1090 |
+
"grad_norm": 72.65499114990234,
|
1091 |
+
"learning_rate": 1e-05,
|
1092 |
+
"loss": 0.1361,
|
1093 |
+
"step": 1500
|
1094 |
+
},
|
1095 |
+
{
|
1096 |
+
"epoch": 32.29946524064171,
|
1097 |
+
"grad_norm": 14.370610237121582,
|
1098 |
+
"learning_rate": 1e-05,
|
1099 |
+
"loss": 0.0346,
|
1100 |
+
"step": 1510
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 32.51336898395722,
|
1104 |
+
"grad_norm": 20.988916397094727,
|
1105 |
+
"learning_rate": 1e-05,
|
1106 |
+
"loss": 0.0025,
|
1107 |
+
"step": 1520
|
1108 |
+
},
|
1109 |
+
{
|
1110 |
+
"epoch": 32.72727272727273,
|
1111 |
+
"grad_norm": 0.002341994782909751,
|
1112 |
+
"learning_rate": 1e-05,
|
1113 |
+
"loss": 0.0004,
|
1114 |
+
"step": 1530
|
1115 |
+
},
|
1116 |
+
{
|
1117 |
+
"epoch": 32.94117647058823,
|
1118 |
+
"grad_norm": 2.404449701309204,
|
1119 |
+
"learning_rate": 1e-05,
|
1120 |
+
"loss": 0.013,
|
1121 |
+
"step": 1540
|
1122 |
+
},
|
1123 |
+
{
|
1124 |
+
"epoch": 33.155080213903744,
|
1125 |
+
"grad_norm": 0.005042242351919413,
|
1126 |
+
"learning_rate": 1e-05,
|
1127 |
+
"loss": 0.0011,
|
1128 |
+
"step": 1550
|
1129 |
+
},
|
1130 |
+
{
|
1131 |
+
"epoch": 33.36898395721925,
|
1132 |
+
"grad_norm": 0.00044189911568537354,
|
1133 |
+
"learning_rate": 1e-05,
|
1134 |
+
"loss": 0.0139,
|
1135 |
+
"step": 1560
|
1136 |
+
},
|
1137 |
+
{
|
1138 |
+
"epoch": 33.58288770053476,
|
1139 |
+
"grad_norm": 0.0014409746509045362,
|
1140 |
+
"learning_rate": 1e-05,
|
1141 |
+
"loss": 0.0284,
|
1142 |
+
"step": 1570
|
1143 |
+
},
|
1144 |
+
{
|
1145 |
+
"epoch": 33.79679144385027,
|
1146 |
+
"grad_norm": 2.608125686645508,
|
1147 |
+
"learning_rate": 1e-05,
|
1148 |
+
"loss": 0.015,
|
1149 |
+
"step": 1580
|
1150 |
+
},
|
1151 |
+
{
|
1152 |
+
"epoch": 34.01069518716577,
|
1153 |
+
"grad_norm": 0.0005788679700344801,
|
1154 |
+
"learning_rate": 1e-05,
|
1155 |
+
"loss": 0.0201,
|
1156 |
+
"step": 1590
|
1157 |
+
},
|
1158 |
+
{
|
1159 |
+
"epoch": 34.224598930481285,
|
1160 |
+
"grad_norm": 75.40068817138672,
|
1161 |
+
"learning_rate": 1e-05,
|
1162 |
+
"loss": 0.028,
|
1163 |
+
"step": 1600
|
1164 |
+
},
|
1165 |
+
{
|
1166 |
+
"epoch": 34.224598930481285,
|
1167 |
+
"eval_accuracy": 0.7,
|
1168 |
+
"eval_loss": 6.69921875,
|
1169 |
+
"eval_runtime": 0.8667,
|
1170 |
+
"eval_samples_per_second": 11.537,
|
1171 |
+
"eval_steps_per_second": 1.154,
|
1172 |
+
"step": 1600
|
1173 |
+
},
|
1174 |
+
{
|
1175 |
+
"epoch": 34.43850267379679,
|
1176 |
+
"grad_norm": 0.005270595662295818,
|
1177 |
+
"learning_rate": 1e-05,
|
1178 |
+
"loss": 0.0087,
|
1179 |
+
"step": 1610
|
1180 |
+
},
|
1181 |
+
{
|
1182 |
+
"epoch": 34.6524064171123,
|
1183 |
+
"grad_norm": 5.715099334716797,
|
1184 |
+
"learning_rate": 1e-05,
|
1185 |
+
"loss": 0.0131,
|
1186 |
+
"step": 1620
|
1187 |
+
},
|
1188 |
+
{
|
1189 |
+
"epoch": 34.86631016042781,
|
1190 |
+
"grad_norm": 7.6973066329956055,
|
1191 |
+
"learning_rate": 1e-05,
|
1192 |
+
"loss": 0.0071,
|
1193 |
+
"step": 1630
|
1194 |
+
},
|
1195 |
+
{
|
1196 |
+
"epoch": 35.080213903743314,
|
1197 |
+
"grad_norm": 0.04163156822323799,
|
1198 |
+
"learning_rate": 1e-05,
|
1199 |
+
"loss": 0.0432,
|
1200 |
+
"step": 1640
|
1201 |
+
},
|
1202 |
+
{
|
1203 |
+
"epoch": 35.294117647058826,
|
1204 |
+
"grad_norm": 0.8603066205978394,
|
1205 |
+
"learning_rate": 1e-05,
|
1206 |
+
"loss": 0.081,
|
1207 |
+
"step": 1650
|
1208 |
+
},
|
1209 |
+
{
|
1210 |
+
"epoch": 35.50802139037433,
|
1211 |
+
"grad_norm": 0.02197711355984211,
|
1212 |
+
"learning_rate": 1e-05,
|
1213 |
+
"loss": 0.0087,
|
1214 |
+
"step": 1660
|
1215 |
+
},
|
1216 |
+
{
|
1217 |
+
"epoch": 35.72192513368984,
|
1218 |
+
"grad_norm": 0.2183353304862976,
|
1219 |
+
"learning_rate": 1e-05,
|
1220 |
+
"loss": 0.0463,
|
1221 |
+
"step": 1670
|
1222 |
+
},
|
1223 |
+
{
|
1224 |
+
"epoch": 35.93582887700535,
|
1225 |
+
"grad_norm": 1.1545729637145996,
|
1226 |
+
"learning_rate": 1e-05,
|
1227 |
+
"loss": 0.0073,
|
1228 |
+
"step": 1680
|
1229 |
+
},
|
1230 |
+
{
|
1231 |
+
"epoch": 36.149732620320854,
|
1232 |
+
"grad_norm": 0.5573898553848267,
|
1233 |
+
"learning_rate": 1e-05,
|
1234 |
+
"loss": 0.0011,
|
1235 |
+
"step": 1690
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 36.36363636363637,
|
1239 |
+
"grad_norm": 0.0014353781007230282,
|
1240 |
+
"learning_rate": 1e-05,
|
1241 |
+
"loss": 0.077,
|
1242 |
+
"step": 1700
|
1243 |
+
},
|
1244 |
+
{
|
1245 |
+
"epoch": 36.57754010695187,
|
1246 |
+
"grad_norm": 0.0148012051358819,
|
1247 |
+
"learning_rate": 1e-05,
|
1248 |
+
"loss": 0.0052,
|
1249 |
+
"step": 1710
|
1250 |
+
},
|
1251 |
+
{
|
1252 |
+
"epoch": 36.79144385026738,
|
1253 |
+
"grad_norm": 0.03192583844065666,
|
1254 |
+
"learning_rate": 1e-05,
|
1255 |
+
"loss": 0.086,
|
1256 |
+
"step": 1720
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 37.00534759358289,
|
1260 |
+
"grad_norm": 82.85762023925781,
|
1261 |
+
"learning_rate": 1e-05,
|
1262 |
+
"loss": 0.0983,
|
1263 |
+
"step": 1730
|
1264 |
+
},
|
1265 |
+
{
|
1266 |
+
"epoch": 37.219251336898395,
|
1267 |
+
"grad_norm": 0.08266828954219818,
|
1268 |
+
"learning_rate": 1e-05,
|
1269 |
+
"loss": 0.0279,
|
1270 |
+
"step": 1740
|
1271 |
+
},
|
1272 |
+
{
|
1273 |
+
"epoch": 37.4331550802139,
|
1274 |
+
"grad_norm": 0.008177646435797215,
|
1275 |
+
"learning_rate": 1e-05,
|
1276 |
+
"loss": 0.0143,
|
1277 |
+
"step": 1750
|
1278 |
+
},
|
1279 |
+
{
|
1280 |
+
"epoch": 37.64705882352941,
|
1281 |
+
"grad_norm": 0.009183383546769619,
|
1282 |
+
"learning_rate": 1e-05,
|
1283 |
+
"loss": 0.0107,
|
1284 |
+
"step": 1760
|
1285 |
+
},
|
1286 |
+
{
|
1287 |
+
"epoch": 37.86096256684492,
|
1288 |
+
"grad_norm": 1.8703384399414062,
|
1289 |
+
"learning_rate": 1e-05,
|
1290 |
+
"loss": 0.0479,
|
1291 |
+
"step": 1770
|
1292 |
+
},
|
1293 |
+
{
|
1294 |
+
"epoch": 38.07486631016043,
|
1295 |
+
"grad_norm": 0.004155300557613373,
|
1296 |
+
"learning_rate": 1e-05,
|
1297 |
+
"loss": 0.0595,
|
1298 |
+
"step": 1780
|
1299 |
+
},
|
1300 |
+
{
|
1301 |
+
"epoch": 38.288770053475936,
|
1302 |
+
"grad_norm": 2.2654850482940674,
|
1303 |
+
"learning_rate": 1e-05,
|
1304 |
+
"loss": 0.033,
|
1305 |
+
"step": 1790
|
1306 |
+
},
|
1307 |
+
{
|
1308 |
+
"epoch": 38.50267379679144,
|
1309 |
+
"grad_norm": 0.20017670094966888,
|
1310 |
+
"learning_rate": 1e-05,
|
1311 |
+
"loss": 0.0078,
|
1312 |
+
"step": 1800
|
1313 |
+
},
|
1314 |
+
{
|
1315 |
+
"epoch": 38.716577540106954,
|
1316 |
+
"grad_norm": 53.791812896728516,
|
1317 |
+
"learning_rate": 1e-05,
|
1318 |
+
"loss": 0.058,
|
1319 |
+
"step": 1810
|
1320 |
+
},
|
1321 |
+
{
|
1322 |
+
"epoch": 38.93048128342246,
|
1323 |
+
"grad_norm": 47.128692626953125,
|
1324 |
+
"learning_rate": 1e-05,
|
1325 |
+
"loss": 0.045,
|
1326 |
+
"step": 1820
|
1327 |
+
},
|
1328 |
+
{
|
1329 |
+
"epoch": 39.144385026737964,
|
1330 |
+
"grad_norm": 41.79620361328125,
|
1331 |
+
"learning_rate": 1e-05,
|
1332 |
+
"loss": 0.156,
|
1333 |
+
"step": 1830
|
1334 |
+
},
|
1335 |
+
{
|
1336 |
+
"epoch": 39.35828877005348,
|
1337 |
+
"grad_norm": 4.738333702087402,
|
1338 |
+
"learning_rate": 1e-05,
|
1339 |
+
"loss": 0.0736,
|
1340 |
+
"step": 1840
|
1341 |
+
},
|
1342 |
+
{
|
1343 |
+
"epoch": 39.57219251336898,
|
1344 |
+
"grad_norm": 0.09291240572929382,
|
1345 |
+
"learning_rate": 1e-05,
|
1346 |
+
"loss": 0.0921,
|
1347 |
+
"step": 1850
|
1348 |
+
},
|
1349 |
+
{
|
1350 |
+
"epoch": 39.786096256684495,
|
1351 |
+
"grad_norm": 11.098824501037598,
|
1352 |
+
"learning_rate": 1e-05,
|
1353 |
+
"loss": 0.0778,
|
1354 |
+
"step": 1860
|
1355 |
+
},
|
1356 |
+
{
|
1357 |
+
"epoch": 40.0,
|
1358 |
+
"grad_norm": 40.247833251953125,
|
1359 |
+
"learning_rate": 1e-05,
|
1360 |
+
"loss": 0.0217,
|
1361 |
+
"step": 1870
|
1362 |
+
},
|
1363 |
+
{
|
1364 |
+
"epoch": 40.213903743315505,
|
1365 |
+
"grad_norm": 2.459455728530884,
|
1366 |
+
"learning_rate": 1e-05,
|
1367 |
+
"loss": 0.008,
|
1368 |
+
"step": 1880
|
1369 |
+
},
|
1370 |
+
{
|
1371 |
+
"epoch": 40.42780748663102,
|
1372 |
+
"grad_norm": 0.03633593022823334,
|
1373 |
+
"learning_rate": 1e-05,
|
1374 |
+
"loss": 0.0053,
|
1375 |
+
"step": 1890
|
1376 |
+
},
|
1377 |
+
{
|
1378 |
+
"epoch": 40.64171122994652,
|
1379 |
+
"grad_norm": 0.0483514666557312,
|
1380 |
+
"learning_rate": 1e-05,
|
1381 |
+
"loss": 0.0566,
|
1382 |
+
"step": 1900
|
1383 |
+
},
|
1384 |
+
{
|
1385 |
+
"epoch": 40.855614973262036,
|
1386 |
+
"grad_norm": 53.63500213623047,
|
1387 |
+
"learning_rate": 1e-05,
|
1388 |
+
"loss": 0.0183,
|
1389 |
+
"step": 1910
|
1390 |
+
},
|
1391 |
+
{
|
1392 |
+
"epoch": 41.06951871657754,
|
1393 |
+
"grad_norm": 0.13120533525943756,
|
1394 |
+
"learning_rate": 1e-05,
|
1395 |
+
"loss": 0.0099,
|
1396 |
+
"step": 1920
|
1397 |
+
},
|
1398 |
+
{
|
1399 |
+
"epoch": 41.283422459893046,
|
1400 |
+
"grad_norm": 0.0008334398153237998,
|
1401 |
+
"learning_rate": 1e-05,
|
1402 |
+
"loss": 0.0007,
|
1403 |
+
"step": 1930
|
1404 |
+
},
|
1405 |
+
{
|
1406 |
+
"epoch": 41.49732620320856,
|
1407 |
+
"grad_norm": 0.028349481523036957,
|
1408 |
+
"learning_rate": 1e-05,
|
1409 |
+
"loss": 0.0011,
|
1410 |
+
"step": 1940
|
1411 |
+
},
|
1412 |
+
{
|
1413 |
+
"epoch": 41.711229946524064,
|
1414 |
+
"grad_norm": 72.34516906738281,
|
1415 |
+
"learning_rate": 1e-05,
|
1416 |
+
"loss": 0.0117,
|
1417 |
+
"step": 1950
|
1418 |
+
},
|
1419 |
+
{
|
1420 |
+
"epoch": 41.92513368983957,
|
1421 |
+
"grad_norm": 0.0020225769840180874,
|
1422 |
+
"learning_rate": 1e-05,
|
1423 |
+
"loss": 0.0097,
|
1424 |
+
"step": 1960
|
1425 |
+
},
|
1426 |
+
{
|
1427 |
+
"epoch": 42.13903743315508,
|
1428 |
+
"grad_norm": 43.7913703918457,
|
1429 |
+
"learning_rate": 1e-05,
|
1430 |
+
"loss": 0.0821,
|
1431 |
+
"step": 1970
|
1432 |
+
},
|
1433 |
+
{
|
1434 |
+
"epoch": 42.35294117647059,
|
1435 |
+
"grad_norm": 18.823827743530273,
|
1436 |
+
"learning_rate": 1e-05,
|
1437 |
+
"loss": 0.01,
|
1438 |
+
"step": 1980
|
1439 |
+
},
|
1440 |
+
{
|
1441 |
+
"epoch": 42.5668449197861,
|
1442 |
+
"grad_norm": 0.01237189769744873,
|
1443 |
+
"learning_rate": 1e-05,
|
1444 |
+
"loss": 0.013,
|
1445 |
+
"step": 1990
|
1446 |
+
},
|
1447 |
+
{
|
1448 |
+
"epoch": 42.780748663101605,
|
1449 |
+
"grad_norm": 0.00256702140904963,
|
1450 |
+
"learning_rate": 1e-05,
|
1451 |
+
"loss": 0.0041,
|
1452 |
+
"step": 2000
|
1453 |
+
},
|
1454 |
+
{
|
1455 |
+
"epoch": 42.780748663101605,
|
1456 |
+
"eval_accuracy": 0.4,
|
1457 |
+
"eval_loss": 7.45703125,
|
1458 |
+
"eval_runtime": 0.871,
|
1459 |
+
"eval_samples_per_second": 11.481,
|
1460 |
+
"eval_steps_per_second": 1.148,
|
1461 |
+
"step": 2000
|
1462 |
+
}
|
1463 |
+
],
|
1464 |
+
"logging_steps": 10,
|
1465 |
+
"max_steps": 2500,
|
1466 |
+
"num_input_tokens_seen": 0,
|
1467 |
+
"num_train_epochs": 55,
|
1468 |
+
"save_steps": 400,
|
1469 |
+
"stateful_callbacks": {
|
1470 |
+
"TrainerControl": {
|
1471 |
+
"args": {
|
1472 |
+
"should_epoch_stop": false,
|
1473 |
+
"should_evaluate": false,
|
1474 |
+
"should_log": false,
|
1475 |
+
"should_save": true,
|
1476 |
+
"should_training_stop": false
|
1477 |
+
},
|
1478 |
+
"attributes": {}
|
1479 |
+
}
|
1480 |
+
},
|
1481 |
+
"total_flos": 1.6510938875124777e+18,
|
1482 |
+
"train_batch_size": 4,
|
1483 |
+
"trial_name": null,
|
1484 |
+
"trial_params": null
|
1485 |
+
}
|
checkpoint-2000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7ea5f51f632ffed5e26cc809a24c0a64055c95fd6a4bba189c05a309473a06bf
|
3 |
+
size 6136
|
checkpoint-2000/zero_to_fp32.py
ADDED
@@ -0,0 +1,604 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/usr/bin/env python
|
2 |
+
|
3 |
+
# Copyright (c) Microsoft Corporation.
|
4 |
+
# SPDX-License-Identifier: Apache-2.0
|
5 |
+
|
6 |
+
# DeepSpeed Team
|
7 |
+
|
8 |
+
# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
|
9 |
+
# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
|
10 |
+
# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
|
11 |
+
# application.
|
12 |
+
#
|
13 |
+
# example: python zero_to_fp32.py . pytorch_model.bin
|
14 |
+
|
15 |
+
import argparse
|
16 |
+
import torch
|
17 |
+
import glob
|
18 |
+
import math
|
19 |
+
import os
|
20 |
+
import re
|
21 |
+
from collections import OrderedDict
|
22 |
+
from dataclasses import dataclass
|
23 |
+
|
24 |
+
# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
|
25 |
+
# DeepSpeed data structures it has to be available in the current python environment.
|
26 |
+
from deepspeed.utils import logger
|
27 |
+
from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
|
28 |
+
FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
|
29 |
+
FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
|
30 |
+
|
31 |
+
|
32 |
+
@dataclass
|
33 |
+
class zero_model_state:
|
34 |
+
buffers: dict()
|
35 |
+
param_shapes: dict()
|
36 |
+
shared_params: list
|
37 |
+
ds_version: int
|
38 |
+
frozen_param_shapes: dict()
|
39 |
+
frozen_param_fragments: dict()
|
40 |
+
|
41 |
+
|
42 |
+
debug = 0
|
43 |
+
|
44 |
+
# load to cpu
|
45 |
+
device = torch.device('cpu')
|
46 |
+
|
47 |
+
|
48 |
+
def atoi(text):
|
49 |
+
return int(text) if text.isdigit() else text
|
50 |
+
|
51 |
+
|
52 |
+
def natural_keys(text):
|
53 |
+
'''
|
54 |
+
alist.sort(key=natural_keys) sorts in human order
|
55 |
+
http://nedbatchelder.com/blog/200712/human_sorting.html
|
56 |
+
(See Toothy's implementation in the comments)
|
57 |
+
'''
|
58 |
+
return [atoi(c) for c in re.split(r'(\d+)', text)]
|
59 |
+
|
60 |
+
|
61 |
+
def get_model_state_file(checkpoint_dir, zero_stage):
|
62 |
+
if not os.path.isdir(checkpoint_dir):
|
63 |
+
raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
|
64 |
+
|
65 |
+
# there should be only one file
|
66 |
+
if zero_stage <= 2:
|
67 |
+
file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
|
68 |
+
elif zero_stage == 3:
|
69 |
+
file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
|
70 |
+
|
71 |
+
if not os.path.exists(file):
|
72 |
+
raise FileNotFoundError(f"can't find model states file at '{file}'")
|
73 |
+
|
74 |
+
return file
|
75 |
+
|
76 |
+
|
77 |
+
def get_checkpoint_files(checkpoint_dir, glob_pattern):
|
78 |
+
# XXX: need to test that this simple glob rule works for multi-node setup too
|
79 |
+
ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
|
80 |
+
|
81 |
+
if len(ckpt_files) == 0:
|
82 |
+
raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
|
83 |
+
|
84 |
+
return ckpt_files
|
85 |
+
|
86 |
+
|
87 |
+
def get_optim_files(checkpoint_dir):
|
88 |
+
return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
|
89 |
+
|
90 |
+
|
91 |
+
def get_model_state_files(checkpoint_dir):
|
92 |
+
return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
|
93 |
+
|
94 |
+
|
95 |
+
def parse_model_states(files):
|
96 |
+
zero_model_states = []
|
97 |
+
for file in files:
|
98 |
+
state_dict = torch.load(file, map_location=device)
|
99 |
+
|
100 |
+
if BUFFER_NAMES not in state_dict:
|
101 |
+
raise ValueError(f"{file} is not a model state checkpoint")
|
102 |
+
buffer_names = state_dict[BUFFER_NAMES]
|
103 |
+
if debug:
|
104 |
+
print("Found buffers:", buffer_names)
|
105 |
+
|
106 |
+
# recover just the buffers while restoring them to fp32 if they were saved in fp16
|
107 |
+
buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
|
108 |
+
param_shapes = state_dict[PARAM_SHAPES]
|
109 |
+
|
110 |
+
# collect parameters that are included in param_shapes
|
111 |
+
param_names = []
|
112 |
+
for s in param_shapes:
|
113 |
+
for name in s.keys():
|
114 |
+
param_names.append(name)
|
115 |
+
|
116 |
+
# update with frozen parameters
|
117 |
+
frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
|
118 |
+
if frozen_param_shapes is not None:
|
119 |
+
if debug:
|
120 |
+
print(f"Found frozen_param_shapes: {frozen_param_shapes}")
|
121 |
+
param_names += list(frozen_param_shapes.keys())
|
122 |
+
|
123 |
+
# handle shared params
|
124 |
+
shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
|
125 |
+
|
126 |
+
ds_version = state_dict.get(DS_VERSION, None)
|
127 |
+
|
128 |
+
frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
|
129 |
+
|
130 |
+
z_model_state = zero_model_state(buffers=buffers,
|
131 |
+
param_shapes=param_shapes,
|
132 |
+
shared_params=shared_params,
|
133 |
+
ds_version=ds_version,
|
134 |
+
frozen_param_shapes=frozen_param_shapes,
|
135 |
+
frozen_param_fragments=frozen_param_fragments)
|
136 |
+
zero_model_states.append(z_model_state)
|
137 |
+
|
138 |
+
return zero_model_states
|
139 |
+
|
140 |
+
|
141 |
+
def parse_optim_states(files, ds_checkpoint_dir):
|
142 |
+
|
143 |
+
total_files = len(files)
|
144 |
+
state_dicts = []
|
145 |
+
for f in files:
|
146 |
+
state_dict = torch.load(f, map_location=device)
|
147 |
+
# immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
|
148 |
+
# and also handle the case where it was already removed by another helper script
|
149 |
+
state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
|
150 |
+
state_dicts.append(state_dict)
|
151 |
+
|
152 |
+
if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
|
153 |
+
raise ValueError(f"{files[0]} is not a zero checkpoint")
|
154 |
+
zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
|
155 |
+
world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
|
156 |
+
|
157 |
+
# For ZeRO-2 each param group can have different partition_count as data parallelism for expert
|
158 |
+
# parameters can be different from data parallelism for non-expert parameters. So we can just
|
159 |
+
# use the max of the partition_count to get the dp world_size.
|
160 |
+
|
161 |
+
if type(world_size) is list:
|
162 |
+
world_size = max(world_size)
|
163 |
+
|
164 |
+
if world_size != total_files:
|
165 |
+
raise ValueError(
|
166 |
+
f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
|
167 |
+
"Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
|
168 |
+
)
|
169 |
+
|
170 |
+
# the groups are named differently in each stage
|
171 |
+
if zero_stage <= 2:
|
172 |
+
fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
|
173 |
+
elif zero_stage == 3:
|
174 |
+
fp32_groups_key = FP32_FLAT_GROUPS
|
175 |
+
else:
|
176 |
+
raise ValueError(f"unknown zero stage {zero_stage}")
|
177 |
+
|
178 |
+
if zero_stage <= 2:
|
179 |
+
fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
|
180 |
+
elif zero_stage == 3:
|
181 |
+
# if there is more than one param group, there will be multiple flattened tensors - one
|
182 |
+
# flattened tensor per group - for simplicity merge them into a single tensor
|
183 |
+
#
|
184 |
+
# XXX: could make the script more memory efficient for when there are multiple groups - it
|
185 |
+
# will require matching the sub-lists of param_shapes for each param group flattened tensor
|
186 |
+
|
187 |
+
fp32_flat_groups = [
|
188 |
+
torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
|
189 |
+
]
|
190 |
+
|
191 |
+
return zero_stage, world_size, fp32_flat_groups
|
192 |
+
|
193 |
+
|
194 |
+
def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
|
195 |
+
"""
|
196 |
+
Returns fp32 state_dict reconstructed from ds checkpoint
|
197 |
+
|
198 |
+
Args:
|
199 |
+
- ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
|
200 |
+
|
201 |
+
"""
|
202 |
+
print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
|
203 |
+
|
204 |
+
optim_files = get_optim_files(ds_checkpoint_dir)
|
205 |
+
zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
|
206 |
+
print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
|
207 |
+
|
208 |
+
model_files = get_model_state_files(ds_checkpoint_dir)
|
209 |
+
|
210 |
+
zero_model_states = parse_model_states(model_files)
|
211 |
+
print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
|
212 |
+
|
213 |
+
if zero_stage <= 2:
|
214 |
+
return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
215 |
+
exclude_frozen_parameters)
|
216 |
+
elif zero_stage == 3:
|
217 |
+
return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
218 |
+
exclude_frozen_parameters)
|
219 |
+
|
220 |
+
|
221 |
+
def _zero2_merge_frozen_params(state_dict, zero_model_states):
|
222 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
223 |
+
return
|
224 |
+
|
225 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
226 |
+
frozen_param_fragments = zero_model_states[0].frozen_param_fragments
|
227 |
+
|
228 |
+
if debug:
|
229 |
+
num_elem = sum(s.numel() for s in frozen_param_shapes.values())
|
230 |
+
print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
231 |
+
|
232 |
+
wanted_params = len(frozen_param_shapes)
|
233 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
234 |
+
avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
|
235 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
236 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
237 |
+
|
238 |
+
total_params = 0
|
239 |
+
total_numel = 0
|
240 |
+
for name, shape in frozen_param_shapes.items():
|
241 |
+
total_params += 1
|
242 |
+
unpartitioned_numel = shape.numel()
|
243 |
+
total_numel += unpartitioned_numel
|
244 |
+
|
245 |
+
state_dict[name] = frozen_param_fragments[name]
|
246 |
+
|
247 |
+
if debug:
|
248 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
249 |
+
|
250 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
251 |
+
|
252 |
+
|
253 |
+
def _has_callable(obj, fn):
|
254 |
+
attr = getattr(obj, fn, None)
|
255 |
+
return callable(attr)
|
256 |
+
|
257 |
+
|
258 |
+
def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
259 |
+
param_shapes = zero_model_states[0].param_shapes
|
260 |
+
|
261 |
+
# Reconstruction protocol:
|
262 |
+
#
|
263 |
+
# XXX: document this
|
264 |
+
|
265 |
+
if debug:
|
266 |
+
for i in range(world_size):
|
267 |
+
for j in range(len(fp32_flat_groups[0])):
|
268 |
+
print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
|
269 |
+
|
270 |
+
# XXX: memory usage doubles here (zero2)
|
271 |
+
num_param_groups = len(fp32_flat_groups[0])
|
272 |
+
merged_single_partition_of_fp32_groups = []
|
273 |
+
for i in range(num_param_groups):
|
274 |
+
merged_partitions = [sd[i] for sd in fp32_flat_groups]
|
275 |
+
full_single_fp32_vector = torch.cat(merged_partitions, 0)
|
276 |
+
merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
|
277 |
+
avail_numel = sum(
|
278 |
+
[full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
|
279 |
+
|
280 |
+
if debug:
|
281 |
+
wanted_params = sum([len(shapes) for shapes in param_shapes])
|
282 |
+
wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
|
283 |
+
# not asserting if there is a mismatch due to possible padding
|
284 |
+
print(f"Have {avail_numel} numels to process.")
|
285 |
+
print(f"Need {wanted_numel} numels in {wanted_params} params.")
|
286 |
+
|
287 |
+
# params
|
288 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
289 |
+
# out-of-core computing solution
|
290 |
+
total_numel = 0
|
291 |
+
total_params = 0
|
292 |
+
for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
|
293 |
+
offset = 0
|
294 |
+
avail_numel = full_single_fp32_vector.numel()
|
295 |
+
for name, shape in shapes.items():
|
296 |
+
|
297 |
+
unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
|
298 |
+
total_numel += unpartitioned_numel
|
299 |
+
total_params += 1
|
300 |
+
|
301 |
+
if debug:
|
302 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
303 |
+
state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
|
304 |
+
offset += unpartitioned_numel
|
305 |
+
|
306 |
+
# Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
|
307 |
+
# avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
|
308 |
+
# paddings performed in the code it's almost impossible to predict the exact numbers w/o the
|
309 |
+
# live optimizer object, so we are checking that the numbers are within the right range
|
310 |
+
align_to = 2 * world_size
|
311 |
+
|
312 |
+
def zero2_align(x):
|
313 |
+
return align_to * math.ceil(x / align_to)
|
314 |
+
|
315 |
+
if debug:
|
316 |
+
print(f"original offset={offset}, avail_numel={avail_numel}")
|
317 |
+
|
318 |
+
offset = zero2_align(offset)
|
319 |
+
avail_numel = zero2_align(avail_numel)
|
320 |
+
|
321 |
+
if debug:
|
322 |
+
print(f"aligned offset={offset}, avail_numel={avail_numel}")
|
323 |
+
|
324 |
+
# Sanity check
|
325 |
+
if offset != avail_numel:
|
326 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
327 |
+
|
328 |
+
print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
|
329 |
+
|
330 |
+
|
331 |
+
def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
332 |
+
exclude_frozen_parameters):
|
333 |
+
state_dict = OrderedDict()
|
334 |
+
|
335 |
+
# buffers
|
336 |
+
buffers = zero_model_states[0].buffers
|
337 |
+
state_dict.update(buffers)
|
338 |
+
if debug:
|
339 |
+
print(f"added {len(buffers)} buffers")
|
340 |
+
|
341 |
+
if not exclude_frozen_parameters:
|
342 |
+
_zero2_merge_frozen_params(state_dict, zero_model_states)
|
343 |
+
|
344 |
+
_zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
345 |
+
|
346 |
+
# recover shared parameters
|
347 |
+
for pair in zero_model_states[0].shared_params:
|
348 |
+
if pair[1] in state_dict:
|
349 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
350 |
+
|
351 |
+
return state_dict
|
352 |
+
|
353 |
+
|
354 |
+
def zero3_partitioned_param_info(unpartitioned_numel, world_size):
|
355 |
+
remainder = unpartitioned_numel % world_size
|
356 |
+
padding_numel = (world_size - remainder) if remainder else 0
|
357 |
+
partitioned_numel = math.ceil(unpartitioned_numel / world_size)
|
358 |
+
return partitioned_numel, padding_numel
|
359 |
+
|
360 |
+
|
361 |
+
def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
|
362 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
363 |
+
return
|
364 |
+
|
365 |
+
if debug:
|
366 |
+
for i in range(world_size):
|
367 |
+
num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
|
368 |
+
print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
369 |
+
|
370 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
371 |
+
wanted_params = len(frozen_param_shapes)
|
372 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
373 |
+
avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
|
374 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
375 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
376 |
+
|
377 |
+
total_params = 0
|
378 |
+
total_numel = 0
|
379 |
+
for name, shape in zero_model_states[0].frozen_param_shapes.items():
|
380 |
+
total_params += 1
|
381 |
+
unpartitioned_numel = shape.numel()
|
382 |
+
total_numel += unpartitioned_numel
|
383 |
+
|
384 |
+
param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
|
385 |
+
state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
|
386 |
+
|
387 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
388 |
+
|
389 |
+
if debug:
|
390 |
+
print(
|
391 |
+
f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
392 |
+
)
|
393 |
+
|
394 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
395 |
+
|
396 |
+
|
397 |
+
def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
398 |
+
param_shapes = zero_model_states[0].param_shapes
|
399 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
400 |
+
# Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
|
401 |
+
# param, re-consolidating each param, while dealing with padding if any
|
402 |
+
|
403 |
+
# merge list of dicts, preserving order
|
404 |
+
param_shapes = {k: v for d in param_shapes for k, v in d.items()}
|
405 |
+
|
406 |
+
if debug:
|
407 |
+
for i in range(world_size):
|
408 |
+
print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
|
409 |
+
|
410 |
+
wanted_params = len(param_shapes)
|
411 |
+
wanted_numel = sum(shape.numel() for shape in param_shapes.values())
|
412 |
+
# not asserting if there is a mismatch due to possible padding
|
413 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
414 |
+
print(f"Trainable params: Have {avail_numel} numels to process.")
|
415 |
+
print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
|
416 |
+
|
417 |
+
# params
|
418 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
419 |
+
# out-of-core computing solution
|
420 |
+
offset = 0
|
421 |
+
total_numel = 0
|
422 |
+
total_params = 0
|
423 |
+
for name, shape in param_shapes.items():
|
424 |
+
|
425 |
+
unpartitioned_numel = shape.numel()
|
426 |
+
total_numel += unpartitioned_numel
|
427 |
+
total_params += 1
|
428 |
+
|
429 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
430 |
+
|
431 |
+
if debug:
|
432 |
+
print(
|
433 |
+
f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
434 |
+
)
|
435 |
+
|
436 |
+
# XXX: memory usage doubles here
|
437 |
+
state_dict[name] = torch.cat(
|
438 |
+
tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
|
439 |
+
0).narrow(0, 0, unpartitioned_numel).view(shape)
|
440 |
+
offset += partitioned_numel
|
441 |
+
|
442 |
+
offset *= world_size
|
443 |
+
|
444 |
+
# Sanity check
|
445 |
+
if offset != avail_numel:
|
446 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
447 |
+
|
448 |
+
print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
|
449 |
+
|
450 |
+
|
451 |
+
def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
452 |
+
exclude_frozen_parameters):
|
453 |
+
state_dict = OrderedDict()
|
454 |
+
|
455 |
+
# buffers
|
456 |
+
buffers = zero_model_states[0].buffers
|
457 |
+
state_dict.update(buffers)
|
458 |
+
if debug:
|
459 |
+
print(f"added {len(buffers)} buffers")
|
460 |
+
|
461 |
+
if not exclude_frozen_parameters:
|
462 |
+
_zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
|
463 |
+
|
464 |
+
_zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
465 |
+
|
466 |
+
# recover shared parameters
|
467 |
+
for pair in zero_model_states[0].shared_params:
|
468 |
+
if pair[1] in state_dict:
|
469 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
470 |
+
|
471 |
+
return state_dict
|
472 |
+
|
473 |
+
|
474 |
+
def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
|
475 |
+
"""
|
476 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
|
477 |
+
``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
|
478 |
+
via a model hub.
|
479 |
+
|
480 |
+
Args:
|
481 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder
|
482 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
|
483 |
+
- ``exclude_frozen_parameters``: exclude frozen parameters
|
484 |
+
|
485 |
+
Returns:
|
486 |
+
- pytorch ``state_dict``
|
487 |
+
|
488 |
+
Note: this approach may not work if your application doesn't have sufficient free CPU memory and
|
489 |
+
you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
|
490 |
+
the checkpoint.
|
491 |
+
|
492 |
+
A typical usage might be ::
|
493 |
+
|
494 |
+
from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
|
495 |
+
# do the training and checkpoint saving
|
496 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
|
497 |
+
model = model.cpu() # move to cpu
|
498 |
+
model.load_state_dict(state_dict)
|
499 |
+
# submit to model hub or save the model to share with others
|
500 |
+
|
501 |
+
In this example the ``model`` will no longer be usable in the deepspeed context of the same
|
502 |
+
application. i.e. you will need to re-initialize the deepspeed engine, since
|
503 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
504 |
+
|
505 |
+
If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
|
506 |
+
|
507 |
+
"""
|
508 |
+
if tag is None:
|
509 |
+
latest_path = os.path.join(checkpoint_dir, 'latest')
|
510 |
+
if os.path.isfile(latest_path):
|
511 |
+
with open(latest_path, 'r') as fd:
|
512 |
+
tag = fd.read().strip()
|
513 |
+
else:
|
514 |
+
raise ValueError(f"Unable to find 'latest' file at {latest_path}")
|
515 |
+
|
516 |
+
ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
|
517 |
+
|
518 |
+
if not os.path.isdir(ds_checkpoint_dir):
|
519 |
+
raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
|
520 |
+
|
521 |
+
return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
|
522 |
+
|
523 |
+
|
524 |
+
def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
|
525 |
+
"""
|
526 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
|
527 |
+
loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
|
528 |
+
|
529 |
+
Args:
|
530 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
531 |
+
- ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
|
532 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
533 |
+
- ``exclude_frozen_parameters``: exclude frozen parameters
|
534 |
+
"""
|
535 |
+
|
536 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
|
537 |
+
print(f"Saving fp32 state dict to {output_file}")
|
538 |
+
torch.save(state_dict, output_file)
|
539 |
+
|
540 |
+
|
541 |
+
def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
|
542 |
+
"""
|
543 |
+
1. Put the provided model to cpu
|
544 |
+
2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
|
545 |
+
3. Load it into the provided model
|
546 |
+
|
547 |
+
Args:
|
548 |
+
- ``model``: the model object to update
|
549 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
550 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
551 |
+
|
552 |
+
Returns:
|
553 |
+
- ``model`: modified model
|
554 |
+
|
555 |
+
Make sure you have plenty of CPU memory available before you call this function. If you don't
|
556 |
+
have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
|
557 |
+
conveniently placed for you in the checkpoint folder.
|
558 |
+
|
559 |
+
A typical usage might be ::
|
560 |
+
|
561 |
+
from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
|
562 |
+
model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
|
563 |
+
# submit to model hub or save the model to share with others
|
564 |
+
|
565 |
+
Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
|
566 |
+
of the same application. i.e. you will need to re-initialize the deepspeed engine, since
|
567 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
568 |
+
|
569 |
+
"""
|
570 |
+
logger.info(f"Extracting fp32 weights")
|
571 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
572 |
+
|
573 |
+
logger.info(f"Overwriting model with fp32 weights")
|
574 |
+
model = model.cpu()
|
575 |
+
model.load_state_dict(state_dict, strict=False)
|
576 |
+
|
577 |
+
return model
|
578 |
+
|
579 |
+
|
580 |
+
if __name__ == "__main__":
|
581 |
+
|
582 |
+
parser = argparse.ArgumentParser()
|
583 |
+
parser.add_argument("checkpoint_dir",
|
584 |
+
type=str,
|
585 |
+
help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
|
586 |
+
parser.add_argument(
|
587 |
+
"output_file",
|
588 |
+
type=str,
|
589 |
+
help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
|
590 |
+
parser.add_argument("-t",
|
591 |
+
"--tag",
|
592 |
+
type=str,
|
593 |
+
default=None,
|
594 |
+
help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
|
595 |
+
parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
|
596 |
+
parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
|
597 |
+
args = parser.parse_args()
|
598 |
+
|
599 |
+
debug = args.debug
|
600 |
+
|
601 |
+
convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
|
602 |
+
args.output_file,
|
603 |
+
tag=args.tag,
|
604 |
+
exclude_frozen_parameters=args.exclude_frozen_parameters)
|