Hugging Face's logo

SynthCode-Project
/
rv2-gemma12b-v4-v4_code_review

Safetensors
Model card Files
xet
 Community
rqadri commited on
Commit
ef3689b
·
verified ·
1 Parent(s): 4361e35

Upload folder using huggingface_hub

Browse files
Files changed (24) hide show
  1. .gitattributes +1 -0
  2. 200M/chat_template.jinja +1 -0
  3. 200M/config.json +170 -0
  4. 200M/generation_config.json +17 -0
  5. 200M/latest +1 -0
  6. 200M/model.safetensors +3 -0
  7. 200M/processor_config.json +62 -0
  8. 200M/rng_state_0.pth +3 -0
  9. 200M/rng_state_1.pth +3 -0
  10. 200M/rng_state_10.pth +3 -0
  11. 200M/rng_state_11.pth +3 -0
  12. 200M/rng_state_2.pth +3 -0
  13. 200M/rng_state_3.pth +3 -0
  14. 200M/rng_state_4.pth +3 -0
  15. 200M/rng_state_5.pth +3 -0
  16. 200M/rng_state_6.pth +3 -0
  17. 200M/rng_state_7.pth +3 -0
  18. 200M/rng_state_8.pth +3 -0
  19. 200M/rng_state_9.pth +3 -0
  20. 200M/scheduler.pt +3 -0
  21. 200M/tokenizer.json +3 -0
  22. 200M/tokenizer_config.json +56 -0
  23. 200M/trainer_state.json +1455 -0
  24. 200M/training_args.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ 200M/tokenizer.json filter=lfs diff=lfs merge=lfs -text
200M/chat_template.jinja ADDED
@@ -0,0 +1 @@
 
 
1
+ {% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content }}{% endif %}{% endfor %}
200M/config.json ADDED
@@ -0,0 +1,170 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Gemma4UnifiedForConditionalGeneration"
4
+ ],
5
+ "audio_config": {
6
+ "_name_or_path": "",
7
+ "architectures": null,
8
+ "audio_embed_dim": 640,
9
+ "chunk_size_feed_forward": 0,
10
+ "dtype": "bfloat16",
11
+ "id2label": {
12
+ "0": "LABEL_0",
13
+ "1": "LABEL_1"
14
+ },
15
+ "initializer_range": 0.02,
16
+ "is_encoder_decoder": false,
17
+ "label2id": {
18
+ "LABEL_0": 0,
19
+ "LABEL_1": 1
20
+ },
21
+ "model_type": "gemma4_unified_audio",
22
+ "output_attentions": false,
23
+ "output_hidden_states": false,
24
+ "problem_type": null,
25
+ "return_dict": true,
26
+ "rms_norm_eps": 1e-06
27
+ },
28
+ "audio_token_id": 258881,
29
+ "boa_token_id": 256000,
30
+ "boi_token_id": 255999,
31
+ "bos_token_id": 2,
32
+ "dtype": "bfloat16",
33
+ "eoa_token_index": 258883,
34
+ "eoi_token_id": 258882,
35
+ "eos_token_id": 1,
36
+ "hidden_size": 3840,
37
+ "image_token_id": 258880,
38
+ "initializer_range": 0.02,
39
+ "model_type": "gemma4_unified",
40
+ "pad_token_id": 0,
41
+ "text_config": {
42
+ "attention_bias": false,
43
+ "attention_dropout": 0.0,
44
+ "attention_k_eq_v": true,
45
+ "bos_token_id": 2,
46
+ "dtype": "bfloat16",
47
+ "enable_moe_block": false,
48
+ "eos_token_id": 1,
49
+ "final_logit_softcapping": 30.0,
50
+ "global_head_dim": 512,
51
+ "head_dim": 256,
52
+ "hidden_activation": "gelu_pytorch_tanh",
53
+ "hidden_size": 3840,
54
+ "hidden_size_per_layer_input": 0,
55
+ "initializer_range": 0.02,
56
+ "intermediate_size": 15360,
57
+ "layer_types": [
58
+ "sliding_attention",
59
+ "sliding_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "sliding_attention",
63
+ "full_attention",
64
+ "sliding_attention",
65
+ "sliding_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "sliding_attention",
69
+ "full_attention",
70
+ "sliding_attention",
71
+ "sliding_attention",
72
+ "sliding_attention",
73
+ "sliding_attention",
74
+ "sliding_attention",
75
+ "full_attention",
76
+ "sliding_attention",
77
+ "sliding_attention",
78
+ "sliding_attention",
79
+ "sliding_attention",
80
+ "sliding_attention",
81
+ "full_attention",
82
+ "sliding_attention",
83
+ "sliding_attention",
84
+ "sliding_attention",
85
+ "sliding_attention",
86
+ "sliding_attention",
87
+ "full_attention",
88
+ "sliding_attention",
89
+ "sliding_attention",
90
+ "sliding_attention",
91
+ "sliding_attention",
92
+ "sliding_attention",
93
+ "full_attention",
94
+ "sliding_attention",
95
+ "sliding_attention",
96
+ "sliding_attention",
97
+ "sliding_attention",
98
+ "sliding_attention",
99
+ "full_attention",
100
+ "sliding_attention",
101
+ "sliding_attention",
102
+ "sliding_attention",
103
+ "sliding_attention",
104
+ "sliding_attention",
105
+ "full_attention"
106
+ ],
107
+ "max_position_embeddings": 262144,
108
+ "model_type": "gemma4_unified_text",
109
+ "moe_intermediate_size": null,
110
+ "num_attention_heads": 16,
111
+ "num_experts": null,
112
+ "num_global_key_value_heads": 1,
113
+ "num_hidden_layers": 48,
114
+ "num_key_value_heads": 8,
115
+ "num_kv_shared_layers": 0,
116
+ "pad_token_id": 0,
117
+ "rms_norm_eps": 1e-06,
118
+ "rope_parameters": {
119
+ "full_attention": {
120
+ "partial_rotary_factor": 0.25,
121
+ "rope_theta": 1000000.0,
122
+ "rope_type": "proportional"
123
+ },
124
+ "sliding_attention": {
125
+ "rope_theta": 10000.0,
126
+ "rope_type": "default"
127
+ }
128
+ },
129
+ "sliding_window": 1024,
130
+ "tie_word_embeddings": true,
131
+ "top_k_experts": null,
132
+ "use_bidirectional_attention": "vision",
133
+ "use_cache": false,
134
+ "use_double_wide_mlp": false,
135
+ "vocab_size": 262144,
136
+ "vocab_size_per_layer_input": 262144
137
+ },
138
+ "tie_word_embeddings": true,
139
+ "transformers_version": "5.10.0",
140
+ "use_cache": false,
141
+ "video_token_id": 258884,
142
+ "vision_config": {
143
+ "_name_or_path": "",
144
+ "architectures": null,
145
+ "chunk_size_feed_forward": 0,
146
+ "dtype": "bfloat16",
147
+ "id2label": {
148
+ "0": "LABEL_0",
149
+ "1": "LABEL_1"
150
+ },
151
+ "initializer_range": 0.02,
152
+ "is_encoder_decoder": false,
153
+ "label2id": {
154
+ "LABEL_0": 0,
155
+ "LABEL_1": 1
156
+ },
157
+ "mm_embed_dim": 3840,
158
+ "mm_posemb_size": 1120,
159
+ "model_type": "gemma4_unified_vision",
160
+ "num_soft_tokens": 280,
161
+ "output_attentions": false,
162
+ "output_hidden_states": false,
163
+ "output_proj_dims": 3840,
164
+ "patch_size": 16,
165
+ "pooling_kernel_size": 3,
166
+ "problem_type": null,
167
+ "return_dict": true,
168
+ "rms_norm_eps": 1e-06
169
+ }
170
+ }
200M/generation_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 2,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 1,
6
+ 1
7
+ ],
8
+ "pad_token_id": 0,
9
+ "suppress_tokens": [
10
+ 258883,
11
+ 258882
12
+ ],
13
+ "temperature": 1.0,
14
+ "top_k": 64,
15
+ "top_p": 0.95,
16
+ "transformers_version": "5.10.0"
17
+ }
200M/latest ADDED
@@ -0,0 +1 @@
 
 
1
+ global_step2035
200M/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e52e2c374ad9a91089a5e3249da1d058097597f8881ffd97708c24ae816d780f
3
+ size 23919549408
200M/processor_config.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audio_ms_per_token": 40,
3
+ "audio_seq_length": 750,
4
+ "feature_extractor": {
5
+ "audio_samples_per_token": 640,
6
+ "feature_extractor_type": "Gemma4UnifiedAudioFeatureExtractor",
7
+ "feature_size": 640,
8
+ "padding_side": "right",
9
+ "padding_value": 0.0,
10
+ "return_attention_mask": true,
11
+ "sampling_rate": 16000
12
+ },
13
+ "image_processor": {
14
+ "do_convert_rgb": true,
15
+ "do_normalize": false,
16
+ "do_rescale": true,
17
+ "do_resize": true,
18
+ "image_mean": [
19
+ 0.0,
20
+ 0.0,
21
+ 0.0
22
+ ],
23
+ "image_processor_type": "Gemma4UnifiedImageProcessor",
24
+ "image_std": [
25
+ 1.0,
26
+ 1.0,
27
+ 1.0
28
+ ],
29
+ "max_soft_tokens": 280,
30
+ "patch_size": 16,
31
+ "pooling_kernel_size": 3,
32
+ "resample": 3,
33
+ "rescale_factor": 0.00392156862745098
34
+ },
35
+ "image_seq_length": 280,
36
+ "processor_class": "Gemma4UnifiedProcessor",
37
+ "video_processor": {
38
+ "do_convert_rgb": true,
39
+ "do_normalize": true,
40
+ "do_rescale": true,
41
+ "do_resize": true,
42
+ "do_sample_frames": true,
43
+ "image_mean": [
44
+ 0.0,
45
+ 0.0,
46
+ 0.0
47
+ ],
48
+ "image_std": [
49
+ 1.0,
50
+ 1.0,
51
+ 1.0
52
+ ],
53
+ "max_soft_tokens": 70,
54
+ "num_frames": 32,
55
+ "patch_size": 16,
56
+ "pooling_kernel_size": 3,
57
+ "resample": 3,
58
+ "rescale_factor": 0.00392156862745098,
59
+ "return_metadata": false,
60
+ "video_processor_type": "Gemma4UnifiedVideoProcessor"
61
+ }
62
+ }
200M/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d7dcf54a8dbd1fff243640ddc034149fbe7c90ed70b8a95a6bb093e989a7577
3
+ size 14469
200M/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2a2b0b9d330140afd742a57df45106276d425d38a78e53b0fd43dba9559f8d4
3
+ size 14469
200M/rng_state_10.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ce222f8f11fd62fd6cb82264500aefbc9e7ef16907ed3e4aa3a8a99c21d95d3
3
+ size 14476
200M/rng_state_11.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d07144bfc4ed675dda70ab66923d17d6f9233560b2db09689cbe1f262c590c88
3
+ size 14476
200M/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5db634f6ed5151d3c898fbe3c365f950516c404bed926de16b253b26e63c7cf
3
+ size 14469
200M/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad774eade44525098fff2d5910e3d16565fca4c2af29ed0bc4d9315a8ac9607a
3
+ size 14469
200M/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebdee1b37be2238764bcfeed9dad268e938248152b0fba60eaf89adda99eac0f
3
+ size 14469
200M/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6046157bd3d69a7cf6ee508b7359b30cb71488a11b11b07a614041b06813c553
3
+ size 14469
200M/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebc9300f101a324d0361f01aec1d57e26b57a703b4e536f4fe982c8ba57a9144
3
+ size 14469
200M/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:357df7b609d6ffd7f22b998281f7856a711ee83505e5ed2835959554b0185df7
3
+ size 14469
200M/rng_state_8.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4fb33667d752a94ade6841f62b8b8fcc22256b4712bd568b141bc690d4a42e11
3
+ size 14469
200M/rng_state_9.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:354a7a38e97ba11b8ad29bf4d00d74f92295b32124cff0ef46baa862aa9b0f2f
3
+ size 14469
200M/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea084231345496123a1e2e3e6e6377891838cfabbf1b42517ab7575e5966916e
3
+ size 1465
200M/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12bac982b793c44b03d52a250a9f0d0b666813da566b910c24a6da0695fd11e6
3
+ size 32170070
200M/tokenizer_config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audio_token": "<|audio|>",
3
+ "backend": "tokenizers",
4
+ "boa_token": "<|audio>",
5
+ "boi_token": "<|image>",
6
+ "bos_token": "<bos>",
7
+ "eoa_token": "<audio|>",
8
+ "eoc_token": "<channel|>",
9
+ "eoi_token": "<image|>",
10
+ "eos_token": "<eos>",
11
+ "eot_token": "<turn|>",
12
+ "escape_token": "<|\"|>",
13
+ "etc_token": "<tool_call|>",
14
+ "etd_token": "<tool|>",
15
+ "etr_token": "<tool_response|>",
16
+ "extra_special_tokens": [
17
+ "<|video|>"
18
+ ],
19
+ "image_token": "<|image|>",
20
+ "is_local": true,
21
+ "local_files_only": true,
22
+ "mask_token": "<mask>",
23
+ "model_max_length": 1000000000000000019884624838656,
24
+ "model_specific_special_tokens": {
25
+ "audio_token": "<|audio|>",
26
+ "boa_token": "<|audio>",
27
+ "boi_token": "<|image>",
28
+ "eoa_token": "<audio|>",
29
+ "eoc_token": "<channel|>",
30
+ "eoi_token": "<image|>",
31
+ "eot_token": "<turn|>",
32
+ "escape_token": "<|\"|>",
33
+ "etc_token": "<tool_call|>",
34
+ "etd_token": "<tool|>",
35
+ "etr_token": "<tool_response|>",
36
+ "image_token": "<|image|>",
37
+ "soc_token": "<|channel>",
38
+ "sot_token": "<|turn>",
39
+ "stc_token": "<|tool_call>",
40
+ "std_token": "<|tool>",
41
+ "str_token": "<|tool_response>",
42
+ "think_token": "<|think|>"
43
+ },
44
+ "pad_token": "<pad>",
45
+ "padding_side": "right",
46
+ "processor_class": "Gemma4UnifiedProcessor",
47
+ "soc_token": "<|channel>",
48
+ "sot_token": "<|turn>",
49
+ "split_special_tokens": false,
50
+ "stc_token": "<|tool_call>",
51
+ "std_token": "<|tool>",
52
+ "str_token": "<|tool_response>",
53
+ "think_token": "<|think|>",
54
+ "tokenizer_class": "GemmaTokenizer",
55
+ "unk_token": "<unk>"
56
+ }
200M/trainer_state.json ADDED
@@ -0,0 +1,1455 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.1097863616745792,
6
+ "eval_steps": 500,
7
+ "global_step": 2035,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0005394907207596029,
14
+ "grad_norm": 369.69301600900053,
15
+ "learning_rate": 2.910112359550562e-06,
16
+ "loss": 1.8730627059936524,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.0010789814415192059,
21
+ "grad_norm": 31.93478853689241,
22
+ "learning_rate": 3.921348314606742e-06,
23
+ "loss": 1.262525463104248,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.0016184721622788088,
28
+ "grad_norm": 141.7556466471986,
29
+ "learning_rate": 4.932584269662922e-06,
30
+ "loss": 1.0669395446777343,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.0021579628830384117,
35
+ "grad_norm": 127.98270782023548,
36
+ "learning_rate": 5.943820224719102e-06,
37
+ "loss": 1.1076365470886231,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.002697453603798015,
42
+ "grad_norm": 101.43748401214339,
43
+ "learning_rate": 6.955056179775282e-06,
44
+ "loss": 1.0109930038452148,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.0032369443245576176,
49
+ "grad_norm": 899.0086783565027,
50
+ "learning_rate": 7.966292134831462e-06,
51
+ "loss": 0.9515115737915039,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.0037764350453172208,
56
+ "grad_norm": 15.557933246692217,
57
+ "learning_rate": 8.977528089887641e-06,
58
+ "loss": 0.9290966987609863,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.0043159257660768235,
63
+ "grad_norm": 256.064896120912,
64
+ "learning_rate": 9.988764044943822e-06,
65
+ "loss": 1.1323322296142577,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.004855416486836427,
70
+ "grad_norm": 20.848525026966247,
71
+ "learning_rate": 1.1000000000000001e-05,
72
+ "loss": 1.088106918334961,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.00539490720759603,
77
+ "grad_norm": 25.60593783327153,
78
+ "learning_rate": 1.2011235955056182e-05,
79
+ "loss": 0.9955007553100585,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.005934397928355632,
84
+ "grad_norm": 11.664176937579581,
85
+ "learning_rate": 1.3022471910112362e-05,
86
+ "loss": 0.942236328125,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.006473888649115235,
91
+ "grad_norm": 7.9248397322900646,
92
+ "learning_rate": 1.403370786516854e-05,
93
+ "loss": 0.9238415718078613,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.007013379369874838,
98
+ "grad_norm": 26.682291952362966,
99
+ "learning_rate": 1.504494382022472e-05,
100
+ "loss": 0.9361964225769043,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.0075528700906344415,
105
+ "grad_norm": 3.6691592434398466,
106
+ "learning_rate": 1.60561797752809e-05,
107
+ "loss": 0.927226448059082,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.008092360811394044,
112
+ "grad_norm": 15.191132047351614,
113
+ "learning_rate": 1.706741573033708e-05,
114
+ "loss": 0.9375665664672852,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.008631851532153647,
119
+ "grad_norm": 7.4209329049996455,
120
+ "learning_rate": 1.8078651685393256e-05,
121
+ "loss": 0.9221304893493653,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.00917134225291325,
126
+ "grad_norm": 5.050234356436225,
127
+ "learning_rate": 1.908988764044944e-05,
128
+ "loss": 0.9327095031738282,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.009710832973672853,
133
+ "grad_norm": 2.532425838265676,
134
+ "learning_rate": 2e-05,
135
+ "loss": 0.90814208984375,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.010250323694432456,
140
+ "grad_norm": 2.8569985447993664,
141
+ "learning_rate": 2e-05,
142
+ "loss": 0.930181884765625,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.01078981441519206,
147
+ "grad_norm": 3.7217733328209746,
148
+ "learning_rate": 2e-05,
149
+ "loss": 0.9137080192565918,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.011329305135951661,
154
+ "grad_norm": 1.6856605125936686,
155
+ "learning_rate": 2e-05,
156
+ "loss": 0.9173044204711914,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.011868795856711264,
161
+ "grad_norm": 4.042279257566151,
162
+ "learning_rate": 2e-05,
163
+ "loss": 0.9044102668762207,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.012408286577470867,
168
+ "grad_norm": 7.693180928481417,
169
+ "learning_rate": 2e-05,
170
+ "loss": 0.9013534545898437,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.01294777729823047,
175
+ "grad_norm": 3.5295771823615647,
176
+ "learning_rate": 2e-05,
177
+ "loss": 0.9148492813110352,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.013487268018990074,
182
+ "grad_norm": 7.839029508109661,
183
+ "learning_rate": 2e-05,
184
+ "loss": 0.9837350845336914,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.014026758739749677,
189
+ "grad_norm": 3.670914514568096,
190
+ "learning_rate": 2e-05,
191
+ "loss": 0.9146605491638183,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.01456624946050928,
196
+ "grad_norm": 3.5577694379027,
197
+ "learning_rate": 2e-05,
198
+ "loss": 0.9401106834411621,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.015105740181268883,
203
+ "grad_norm": 2.453202809896575,
204
+ "learning_rate": 2e-05,
205
+ "loss": 0.9477849960327148,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.015645230902028486,
210
+ "grad_norm": 1.9729075743285769,
211
+ "learning_rate": 2e-05,
212
+ "loss": 0.9066492080688476,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.016184721622788088,
217
+ "grad_norm": 2.8099192390223795,
218
+ "learning_rate": 2e-05,
219
+ "loss": 0.9063838958740235,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.016724212343547692,
224
+ "grad_norm": 4.236623113420927,
225
+ "learning_rate": 2e-05,
226
+ "loss": 0.8637477874755859,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.017263703064307294,
231
+ "grad_norm": 4.167628129961564,
232
+ "learning_rate": 2e-05,
233
+ "loss": 0.9061297416687012,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.017803193785066895,
238
+ "grad_norm": 6.935834245033307,
239
+ "learning_rate": 2e-05,
240
+ "loss": 0.8871530532836914,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.0183426845058265,
245
+ "grad_norm": 3.2423751981314384,
246
+ "learning_rate": 2e-05,
247
+ "loss": 0.9223550796508789,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.0188821752265861,
252
+ "grad_norm": 30.606793216440163,
253
+ "learning_rate": 2e-05,
254
+ "loss": 3.54920654296875,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.019421665947345707,
259
+ "grad_norm": 4.550939608135107,
260
+ "learning_rate": 2e-05,
261
+ "loss": 0.9992364883422852,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.019961156668105308,
266
+ "grad_norm": 10.355788989674835,
267
+ "learning_rate": 2e-05,
268
+ "loss": 0.9463441848754883,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.020500647388864913,
273
+ "grad_norm": 9.470092673566741,
274
+ "learning_rate": 2e-05,
275
+ "loss": 0.9165935516357422,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.021040138109624514,
280
+ "grad_norm": 16.04366173130251,
281
+ "learning_rate": 2e-05,
282
+ "loss": 0.9199161529541016,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.02157962883038412,
287
+ "grad_norm": 16.276111063417574,
288
+ "learning_rate": 2e-05,
289
+ "loss": 0.8900386810302734,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.02211911955114372,
294
+ "grad_norm": 14.653812444903842,
295
+ "learning_rate": 2e-05,
296
+ "loss": 0.8979652404785157,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.022658610271903322,
301
+ "grad_norm": 2.6977548589169094,
302
+ "learning_rate": 2e-05,
303
+ "loss": 0.9128728866577148,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.023198100992662927,
308
+ "grad_norm": 23.582887550971233,
309
+ "learning_rate": 2e-05,
310
+ "loss": 0.8891826629638672,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.023737591713422528,
315
+ "grad_norm": 2.9436146384915443,
316
+ "learning_rate": 2e-05,
317
+ "loss": 0.9091741561889648,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.024277082434182133,
322
+ "grad_norm": 3.8178614247451996,
323
+ "learning_rate": 2e-05,
324
+ "loss": 0.9071731567382812,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.024816573154941735,
329
+ "grad_norm": 9.900017833351434,
330
+ "learning_rate": 2e-05,
331
+ "loss": 0.8803417205810546,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.02535606387570134,
336
+ "grad_norm": 11.377800042408293,
337
+ "learning_rate": 2e-05,
338
+ "loss": 0.9077526092529297,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.02589555459646094,
343
+ "grad_norm": 4.425640907224202,
344
+ "learning_rate": 2e-05,
345
+ "loss": 0.8951313018798828,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.026435045317220542,
350
+ "grad_norm": 5.543696039922662,
351
+ "learning_rate": 2e-05,
352
+ "loss": 0.9170653343200683,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.026974536037980147,
357
+ "grad_norm": 6.637675684831988,
358
+ "learning_rate": 2e-05,
359
+ "loss": 0.9328924179077148,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.02751402675873975,
364
+ "grad_norm": 4.541587002846055,
365
+ "learning_rate": 2e-05,
366
+ "loss": 0.9184404373168945,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.028053517479499353,
371
+ "grad_norm": 4.834955555833834,
372
+ "learning_rate": 2e-05,
373
+ "loss": 0.94840087890625,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.028593008200258955,
378
+ "grad_norm": 2.7476196554801118,
379
+ "learning_rate": 2e-05,
380
+ "loss": 0.9112020492553711,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.02913249892101856,
385
+ "grad_norm": 3.729222437612153,
386
+ "learning_rate": 2e-05,
387
+ "loss": 0.8715265274047852,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.02967198964177816,
392
+ "grad_norm": 1.8986444153489899,
393
+ "learning_rate": 2e-05,
394
+ "loss": 0.8817804336547852,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.030211480362537766,
399
+ "grad_norm": 19.652863197507482,
400
+ "learning_rate": 2e-05,
401
+ "loss": 0.9034578323364257,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.030750971083297367,
406
+ "grad_norm": 4.893935787998432,
407
+ "learning_rate": 2e-05,
408
+ "loss": 0.9201476097106933,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.03129046180405697,
413
+ "grad_norm": 2.121315677688142,
414
+ "learning_rate": 2e-05,
415
+ "loss": 0.9087504386901856,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.031829952524816574,
420
+ "grad_norm": 1.6805303205822455,
421
+ "learning_rate": 2e-05,
422
+ "loss": 0.8938014030456543,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.032369443245576175,
427
+ "grad_norm": 2.091016561123043,
428
+ "learning_rate": 2e-05,
429
+ "loss": 0.8701451301574707,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.03290893396633578,
434
+ "grad_norm": 4.4795444497420345,
435
+ "learning_rate": 2e-05,
436
+ "loss": 0.8988707542419434,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.033448424687095385,
441
+ "grad_norm": 1.40755956780777,
442
+ "learning_rate": 2e-05,
443
+ "loss": 0.8736795425415039,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.033987915407854986,
448
+ "grad_norm": 1.9677540233275401,
449
+ "learning_rate": 2e-05,
450
+ "loss": 0.8994430541992188,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.03452740612861459,
455
+ "grad_norm": 2.8537882081306605,
456
+ "learning_rate": 2e-05,
457
+ "loss": 0.8869138717651367,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.03506689684937419,
462
+ "grad_norm": 2.3406412447640133,
463
+ "learning_rate": 2e-05,
464
+ "loss": 0.8736772537231445,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.03560638757013379,
469
+ "grad_norm": 2.7002297840083647,
470
+ "learning_rate": 2e-05,
471
+ "loss": 0.8878070831298828,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.0361458782908934,
476
+ "grad_norm": 5.478051955992026,
477
+ "learning_rate": 2e-05,
478
+ "loss": 0.9007582664489746,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.036685369011653,
483
+ "grad_norm": 5.987775578117277,
484
+ "learning_rate": 2e-05,
485
+ "loss": 0.877619457244873,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.0372248597324126,
490
+ "grad_norm": 4.062183705280081,
491
+ "learning_rate": 2e-05,
492
+ "loss": 0.9615988731384277,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.0377643504531722,
497
+ "grad_norm": 1.7304116921547956,
498
+ "learning_rate": 2e-05,
499
+ "loss": 0.8999065399169922,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.03830384117393181,
504
+ "grad_norm": 10.224091687714415,
505
+ "learning_rate": 2e-05,
506
+ "loss": 0.8775262832641602,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.03884333189469141,
511
+ "grad_norm": 2.6973024084894885,
512
+ "learning_rate": 2e-05,
513
+ "loss": 0.88863525390625,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.039382822615451014,
518
+ "grad_norm": 3.0097373941657186,
519
+ "learning_rate": 2e-05,
520
+ "loss": 0.8921207427978516,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.039922313336210616,
525
+ "grad_norm": 2.4929780841655207,
526
+ "learning_rate": 2e-05,
527
+ "loss": 0.8870817184448242,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.04046180405697022,
532
+ "grad_norm": 2.680211035148592,
533
+ "learning_rate": 2e-05,
534
+ "loss": 0.8718931198120117,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.041001294777729826,
539
+ "grad_norm": 3.314725765838302,
540
+ "learning_rate": 2e-05,
541
+ "loss": 0.8696352005004883,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.04154078549848943,
546
+ "grad_norm": 2.8067410842808824,
547
+ "learning_rate": 2e-05,
548
+ "loss": 0.8808348655700684,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.04208027621924903,
553
+ "grad_norm": 4.568417784852867,
554
+ "learning_rate": 2e-05,
555
+ "loss": 0.8762812614440918,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.04261976694000863,
560
+ "grad_norm": 1.591466243191371,
561
+ "learning_rate": 2e-05,
562
+ "loss": 0.8897010803222656,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.04315925766076824,
567
+ "grad_norm": 1.7378624546635357,
568
+ "learning_rate": 2e-05,
569
+ "loss": 0.8871335983276367,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.04369874838152784,
574
+ "grad_norm": 5.282884941300159,
575
+ "learning_rate": 2e-05,
576
+ "loss": 0.8862749099731445,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.04423823910228744,
581
+ "grad_norm": 1.4282666163813993,
582
+ "learning_rate": 2e-05,
583
+ "loss": 0.9045479774475098,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.04477772982304704,
588
+ "grad_norm": 1.804678546954577,
589
+ "learning_rate": 2e-05,
590
+ "loss": 0.8669574737548829,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.045317220543806644,
595
+ "grad_norm": 2.7003745816690836,
596
+ "learning_rate": 2e-05,
597
+ "loss": 0.8664716720581055,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.04585671126456625,
602
+ "grad_norm": 2.8347718319860626,
603
+ "learning_rate": 2e-05,
604
+ "loss": 0.8740848541259766,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.046396201985325854,
609
+ "grad_norm": 2.3599301199842655,
610
+ "learning_rate": 2e-05,
611
+ "loss": 0.8661260604858398,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.046935692706085455,
616
+ "grad_norm": 2.675130769028597,
617
+ "learning_rate": 2e-05,
618
+ "loss": 0.8940586090087891,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.047475183426845056,
623
+ "grad_norm": 2.13219637159858,
624
+ "learning_rate": 2e-05,
625
+ "loss": 0.8656101226806641,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.04801467414760466,
630
+ "grad_norm": 5.746935578097955,
631
+ "learning_rate": 2e-05,
632
+ "loss": 0.9049406051635742,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.048554164868364266,
637
+ "grad_norm": 3.040442087390683,
638
+ "learning_rate": 2e-05,
639
+ "loss": 0.8673805236816406,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.04909365558912387,
644
+ "grad_norm": 1.8033864790060448,
645
+ "learning_rate": 2e-05,
646
+ "loss": 0.9076686859130859,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.04963314630988347,
651
+ "grad_norm": 3.7270371618478446,
652
+ "learning_rate": 2e-05,
653
+ "loss": 0.8856944084167481,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.05017263703064307,
658
+ "grad_norm": 18.34533608060895,
659
+ "learning_rate": 2e-05,
660
+ "loss": 0.8803435325622558,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.05071212775140268,
665
+ "grad_norm": 7.031326732375769,
666
+ "learning_rate": 2e-05,
667
+ "loss": 0.8922122955322266,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.05125161847216228,
672
+ "grad_norm": 1.271585738790212,
673
+ "learning_rate": 2e-05,
674
+ "loss": 0.8702299118041992,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.05179110919292188,
679
+ "grad_norm": 3.1087560610639984,
680
+ "learning_rate": 2e-05,
681
+ "loss": 0.869837474822998,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.05233059991368148,
686
+ "grad_norm": 1.336885754811761,
687
+ "learning_rate": 2e-05,
688
+ "loss": 0.8768095016479492,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.052870090634441085,
693
+ "grad_norm": 1.2937187665672654,
694
+ "learning_rate": 2e-05,
695
+ "loss": 0.8737648010253907,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.05340958135520069,
700
+ "grad_norm": 6.496337599049736,
701
+ "learning_rate": 2e-05,
702
+ "loss": 0.8852460861206055,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.053949072075960294,
707
+ "grad_norm": 2.314909026977299,
708
+ "learning_rate": 2e-05,
709
+ "loss": 0.9168342590332031,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.054488562796719896,
714
+ "grad_norm": 1.3797589526126182,
715
+ "learning_rate": 2e-05,
716
+ "loss": 0.9012920379638671,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.0550280535174795,
721
+ "grad_norm": 1.7456393044082175,
722
+ "learning_rate": 2e-05,
723
+ "loss": 0.8839190483093262,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.055567544238239105,
728
+ "grad_norm": 2.2368924316262913,
729
+ "learning_rate": 2e-05,
730
+ "loss": 0.8841041564941406,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.05610703495899871,
735
+ "grad_norm": 2.3060976347623408,
736
+ "learning_rate": 2e-05,
737
+ "loss": 0.9180487632751465,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.05664652567975831,
742
+ "grad_norm": 2.2473033744197655,
743
+ "learning_rate": 2e-05,
744
+ "loss": 0.8983830451965332,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.05718601640051791,
749
+ "grad_norm": 3.33794913543599,
750
+ "learning_rate": 2e-05,
751
+ "loss": 0.9058303833007812,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.05772550712127751,
756
+ "grad_norm": 2.1863704996804874,
757
+ "learning_rate": 2e-05,
758
+ "loss": 0.8500079154968262,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.05826499784203712,
763
+ "grad_norm": 5.464104262184261,
764
+ "learning_rate": 2e-05,
765
+ "loss": 0.86676025390625,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.05880448856279672,
770
+ "grad_norm": 1.2580865472126652,
771
+ "learning_rate": 2e-05,
772
+ "loss": 0.8795917510986329,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.05934397928355632,
777
+ "grad_norm": 1.5392878698728925,
778
+ "learning_rate": 2e-05,
779
+ "loss": 0.855427360534668,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.059883470004315924,
784
+ "grad_norm": 1.6600456111757242,
785
+ "learning_rate": 2e-05,
786
+ "loss": 0.8580084800720215,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.06042296072507553,
791
+ "grad_norm": 1.9728666357091431,
792
+ "learning_rate": 2e-05,
793
+ "loss": 0.8592136383056641,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.060962451445835134,
798
+ "grad_norm": 2.622532245761096,
799
+ "learning_rate": 2e-05,
800
+ "loss": 0.8778785705566406,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.061501942166594735,
805
+ "grad_norm": 2.436300801764917,
806
+ "learning_rate": 2e-05,
807
+ "loss": 0.8558690071105957,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.062041432887354336,
812
+ "grad_norm": 2.072486307736441,
813
+ "learning_rate": 2e-05,
814
+ "loss": 0.8554096221923828,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.06258092360811394,
819
+ "grad_norm": 2.9578754600495105,
820
+ "learning_rate": 2e-05,
821
+ "loss": 0.8688982963562012,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.06312041432887354,
826
+ "grad_norm": 2.2185001887709106,
827
+ "learning_rate": 2e-05,
828
+ "loss": 0.8960395812988281,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.06365990504963315,
833
+ "grad_norm": 2.39317741084652,
834
+ "learning_rate": 2e-05,
835
+ "loss": 0.8809322357177735,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.06419939577039276,
840
+ "grad_norm": 7.053593954669792,
841
+ "learning_rate": 2e-05,
842
+ "loss": 0.8588828086853028,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.06473888649115235,
847
+ "grad_norm": 2.8316276030407392,
848
+ "learning_rate": 2e-05,
849
+ "loss": 0.863832664489746,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.06527837721191196,
854
+ "grad_norm": 55.263846840104094,
855
+ "learning_rate": 2e-05,
856
+ "loss": 0.8467828750610351,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.06581786793267155,
861
+ "grad_norm": 1.891195934618861,
862
+ "learning_rate": 2e-05,
863
+ "loss": 0.8810437202453614,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.06635735865343116,
868
+ "grad_norm": 4.262649131447584,
869
+ "learning_rate": 2e-05,
870
+ "loss": 0.8573444366455079,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.06689684937419077,
875
+ "grad_norm": 5.280398859967369,
876
+ "learning_rate": 2e-05,
877
+ "loss": 0.8775759696960449,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.06743634009495036,
882
+ "grad_norm": 13.161950494331434,
883
+ "learning_rate": 2e-05,
884
+ "loss": 0.8802067756652832,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.06797583081570997,
889
+ "grad_norm": 2.3840227633253708,
890
+ "learning_rate": 2e-05,
891
+ "loss": 0.8745565414428711,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.06851532153646957,
896
+ "grad_norm": 2.578881266139901,
897
+ "learning_rate": 2e-05,
898
+ "loss": 0.8929662704467773,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.06905481225722918,
903
+ "grad_norm": 1.3551093714665399,
904
+ "learning_rate": 2e-05,
905
+ "loss": 0.8856276512145996,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.06959430297798878,
910
+ "grad_norm": 1.421718866393892,
911
+ "learning_rate": 2e-05,
912
+ "loss": 0.8779384613037109,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.07013379369874838,
917
+ "grad_norm": 1.3412253586411367,
918
+ "learning_rate": 2e-05,
919
+ "loss": 0.8418590545654296,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.07067328441950799,
924
+ "grad_norm": 1.3164576570437523,
925
+ "learning_rate": 2e-05,
926
+ "loss": 0.8589597702026367,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.07121277514026758,
931
+ "grad_norm": 5.63391297375713,
932
+ "learning_rate": 2e-05,
933
+ "loss": 0.8794595718383789,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.07175226586102719,
938
+ "grad_norm": 1.3610004564150833,
939
+ "learning_rate": 2e-05,
940
+ "loss": 0.8666675567626954,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.0722917565817868,
945
+ "grad_norm": 1.7200672230946694,
946
+ "learning_rate": 2e-05,
947
+ "loss": 0.8707124710083007,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.07283124730254639,
952
+ "grad_norm": 1.3160601308789588,
953
+ "learning_rate": 2e-05,
954
+ "loss": 0.8457590103149414,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.073370738023306,
959
+ "grad_norm": 1.7778775450463142,
960
+ "learning_rate": 2e-05,
961
+ "loss": 0.8669673919677734,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.0739102287440656,
966
+ "grad_norm": 2.2742353103065485,
967
+ "learning_rate": 2e-05,
968
+ "loss": 0.920014762878418,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.0744497194648252,
973
+ "grad_norm": 4.185305052537614,
974
+ "learning_rate": 2e-05,
975
+ "loss": 0.9066905021667481,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.07498921018558481,
980
+ "grad_norm": 7.045183411349042,
981
+ "learning_rate": 2e-05,
982
+ "loss": 0.8720160484313965,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.0755287009063444,
987
+ "grad_norm": 2.377204425821727,
988
+ "learning_rate": 2e-05,
989
+ "loss": 0.9035792350769043,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.07606819162710401,
994
+ "grad_norm": 1.5514726466038995,
995
+ "learning_rate": 2e-05,
996
+ "loss": 0.8550067901611328,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.07660768234786362,
1001
+ "grad_norm": 3.2194246326766627,
1002
+ "learning_rate": 2e-05,
1003
+ "loss": 0.8613847732543946,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.07714717306862322,
1008
+ "grad_norm": 2.191413489389274,
1009
+ "learning_rate": 2e-05,
1010
+ "loss": 0.8483789443969727,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.07768666378938283,
1015
+ "grad_norm": 2.561083258333701,
1016
+ "learning_rate": 2e-05,
1017
+ "loss": 0.8561199188232422,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.07822615451014242,
1022
+ "grad_norm": 2.799486266864444,
1023
+ "learning_rate": 2e-05,
1024
+ "loss": 0.8749614715576172,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.07876564523090203,
1029
+ "grad_norm": 1.6625068228812598,
1030
+ "learning_rate": 2e-05,
1031
+ "loss": 0.8524051666259765,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.07930513595166164,
1036
+ "grad_norm": 1.5627273498445615,
1037
+ "learning_rate": 2e-05,
1038
+ "loss": 0.8613437652587891,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.07984462667242123,
1043
+ "grad_norm": 3.1187718040693815,
1044
+ "learning_rate": 2e-05,
1045
+ "loss": 0.8489472389221191,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.08038411739318084,
1050
+ "grad_norm": 1.999242227430214,
1051
+ "learning_rate": 2e-05,
1052
+ "loss": 0.8557394027709961,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.08092360811394043,
1057
+ "grad_norm": 2.1662789248263086,
1058
+ "learning_rate": 2e-05,
1059
+ "loss": 0.8410913467407226,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.08146309883470004,
1064
+ "grad_norm": 1.117593071600269,
1065
+ "learning_rate": 2e-05,
1066
+ "loss": 0.8719884872436523,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.08200258955545965,
1071
+ "grad_norm": 1.7282463420500256,
1072
+ "learning_rate": 2e-05,
1073
+ "loss": 0.8429347991943359,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.08254208027621925,
1078
+ "grad_norm": 1.10160222981424,
1079
+ "learning_rate": 2e-05,
1080
+ "loss": 0.8589319229125977,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.08308157099697885,
1085
+ "grad_norm": 1.657648568922712,
1086
+ "learning_rate": 2e-05,
1087
+ "loss": 0.8519350051879883,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.08362106171773845,
1092
+ "grad_norm": 1.2061011251185034,
1093
+ "learning_rate": 2e-05,
1094
+ "loss": 0.8548383712768555,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.08416055243849806,
1099
+ "grad_norm": 1.6503845091419074,
1100
+ "learning_rate": 2e-05,
1101
+ "loss": 0.8504886627197266,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.08470004315925767,
1106
+ "grad_norm": 1.650450748636652,
1107
+ "learning_rate": 2e-05,
1108
+ "loss": 0.8272921562194824,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.08523953388001726,
1113
+ "grad_norm": 1.2153538779996496,
1114
+ "learning_rate": 2e-05,
1115
+ "loss": 0.8586090087890625,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.08577902460077687,
1120
+ "grad_norm": 1.4119654167230689,
1121
+ "learning_rate": 2e-05,
1122
+ "loss": 0.8150126457214355,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.08631851532153648,
1127
+ "grad_norm": 2.0481812733972435,
1128
+ "learning_rate": 2e-05,
1129
+ "loss": 0.8605268478393555,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.08685800604229607,
1134
+ "grad_norm": 1.3493621359452488,
1135
+ "learning_rate": 2e-05,
1136
+ "loss": 0.8761652946472168,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.08739749676305568,
1141
+ "grad_norm": 1.937128776151837,
1142
+ "learning_rate": 2e-05,
1143
+ "loss": 0.8717695236206054,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.08793698748381527,
1148
+ "grad_norm": 1.2331164884742856,
1149
+ "learning_rate": 2e-05,
1150
+ "loss": 0.8695238113403321,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.08847647820457488,
1155
+ "grad_norm": 1.0766868462241113,
1156
+ "learning_rate": 2e-05,
1157
+ "loss": 0.8439886093139648,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.08901596892533449,
1162
+ "grad_norm": 4.825188821823706,
1163
+ "learning_rate": 2e-05,
1164
+ "loss": 0.8383602142333985,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.08955545964609408,
1169
+ "grad_norm": 1.0473009888263762,
1170
+ "learning_rate": 2e-05,
1171
+ "loss": 0.8446158409118653,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.0900949503668537,
1176
+ "grad_norm": 2.1550692634060407,
1177
+ "learning_rate": 2e-05,
1178
+ "loss": 0.8526595115661622,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.09063444108761329,
1183
+ "grad_norm": 1.1805712525240586,
1184
+ "learning_rate": 2e-05,
1185
+ "loss": 0.8445590019226075,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.0911739318083729,
1190
+ "grad_norm": 1.3154682778410502,
1191
+ "learning_rate": 2e-05,
1192
+ "loss": 0.8289753913879394,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.0917134225291325,
1197
+ "grad_norm": 1.7425671016352309,
1198
+ "learning_rate": 2e-05,
1199
+ "loss": 0.8550616264343261,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.0922529132498921,
1204
+ "grad_norm": 1.809326593705475,
1205
+ "learning_rate": 2e-05,
1206
+ "loss": 0.8592042922973633,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.09279240397065171,
1211
+ "grad_norm": 2.294726987915818,
1212
+ "learning_rate": 2e-05,
1213
+ "loss": 0.8464397430419922,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.0933318946914113,
1218
+ "grad_norm": 1.3341335707857838,
1219
+ "learning_rate": 2e-05,
1220
+ "loss": 0.8401545524597168,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.09387138541217091,
1225
+ "grad_norm": 1.7059856939535623,
1226
+ "learning_rate": 2e-05,
1227
+ "loss": 0.850313663482666,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.09441087613293052,
1232
+ "grad_norm": 1.1913305227905306,
1233
+ "learning_rate": 2e-05,
1234
+ "loss": 0.8246526718139648,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.09495036685369011,
1239
+ "grad_norm": 2.4712263811656063,
1240
+ "learning_rate": 2e-05,
1241
+ "loss": 0.8324945449829102,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.09548985757444972,
1246
+ "grad_norm": 1.2460535060628168,
1247
+ "learning_rate": 2e-05,
1248
+ "loss": 0.83499755859375,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.09602934829520932,
1253
+ "grad_norm": 0.9541283688175325,
1254
+ "learning_rate": 2e-05,
1255
+ "loss": 0.8438259124755859,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.09656883901596892,
1260
+ "grad_norm": 4.15270429876545,
1261
+ "learning_rate": 2e-05,
1262
+ "loss": 0.8374526977539063,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.09710832973672853,
1267
+ "grad_norm": 1.2843301784727401,
1268
+ "learning_rate": 2e-05,
1269
+ "loss": 0.855588722229004,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.09764782045748813,
1274
+ "grad_norm": 1.2852787584009888,
1275
+ "learning_rate": 2e-05,
1276
+ "loss": 0.8446294784545898,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.09818731117824774,
1281
+ "grad_norm": 1.1732513743084148,
1282
+ "learning_rate": 2e-05,
1283
+ "loss": 0.8550775527954102,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.09872680189900734,
1288
+ "grad_norm": 1.0484004988454474,
1289
+ "learning_rate": 2e-05,
1290
+ "loss": 0.8461496353149414,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.09926629261976694,
1295
+ "grad_norm": 1.1796374157652922,
1296
+ "learning_rate": 2e-05,
1297
+ "loss": 0.8495699882507324,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.09980578334052655,
1302
+ "grad_norm": 1.530009834654909,
1303
+ "learning_rate": 2e-05,
1304
+ "loss": 0.8529845237731933,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.10034527406128614,
1309
+ "grad_norm": 1.3821515513567662,
1310
+ "learning_rate": 2e-05,
1311
+ "loss": 0.8383827209472656,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.10088476478204575,
1316
+ "grad_norm": 1.433339075396923,
1317
+ "learning_rate": 2e-05,
1318
+ "loss": 0.8441427230834961,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.10142425550280536,
1323
+ "grad_norm": 1.4202627926244262,
1324
+ "learning_rate": 2e-05,
1325
+ "loss": 0.8355476379394531,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.10196374622356495,
1330
+ "grad_norm": 1.275529436709381,
1331
+ "learning_rate": 2e-05,
1332
+ "loss": 0.8475621223449707,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.10250323694432456,
1337
+ "grad_norm": 4.886608266675648,
1338
+ "learning_rate": 2e-05,
1339
+ "loss": 0.8222907066345215,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.10304272766508416,
1344
+ "grad_norm": 1.1032859405565207,
1345
+ "learning_rate": 2e-05,
1346
+ "loss": 0.8717263221740723,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.10358221838584376,
1351
+ "grad_norm": 1.4792708796316567,
1352
+ "learning_rate": 2e-05,
1353
+ "loss": 0.8422025680541992,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.10412170910660337,
1358
+ "grad_norm": 1.052829235330944,
1359
+ "learning_rate": 2e-05,
1360
+ "loss": 0.8275224685668945,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.10466119982736297,
1365
+ "grad_norm": 1.5243538008576782,
1366
+ "learning_rate": 2e-05,
1367
+ "loss": 0.8201045989990234,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.10520069054812257,
1372
+ "grad_norm": 1.8549508993672652,
1373
+ "learning_rate": 2e-05,
1374
+ "loss": 0.8507599830627441,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.10574018126888217,
1379
+ "grad_norm": 3.2676775863279293,
1380
+ "learning_rate": 2e-05,
1381
+ "loss": 0.8315736770629882,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.10627967198964178,
1386
+ "grad_norm": 3.170505632566329,
1387
+ "learning_rate": 2e-05,
1388
+ "loss": 0.8597011566162109,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.10681916271040139,
1393
+ "grad_norm": 1.225518275743383,
1394
+ "learning_rate": 2e-05,
1395
+ "loss": 0.8449475288391113,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.10735865343116098,
1400
+ "grad_norm": 1.7639666202458668,
1401
+ "learning_rate": 2e-05,
1402
+ "loss": 0.8267492294311524,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.10789814415192059,
1407
+ "grad_norm": 1.7820697092202835,
1408
+ "learning_rate": 2e-05,
1409
+ "loss": 0.8312843322753907,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.10843763487268018,
1414
+ "grad_norm": 4.015953726833389,
1415
+ "learning_rate": 2e-05,
1416
+ "loss": 0.8443691253662109,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.10897712559343979,
1421
+ "grad_norm": 1.999195946930485,
1422
+ "learning_rate": 2e-05,
1423
+ "loss": 0.8289900779724121,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.1095166163141994,
1428
+ "grad_norm": 1.8651265962731969,
1429
+ "learning_rate": 2e-05,
1430
+ "loss": 0.840304946899414,
1431
+ "step": 2030
1432
+ }
1433
+ ],
1434
+ "logging_steps": 10,
1435
+ "max_steps": 5935,
1436
+ "num_input_tokens_seen": 0,
1437
+ "num_train_epochs": 1,
1438
+ "save_steps": 2035,
1439
+ "stateful_callbacks": {
1440
+ "TrainerControl": {
1441
+ "args": {
1442
+ "should_epoch_stop": false,
1443
+ "should_evaluate": false,
1444
+ "should_log": false,
1445
+ "should_save": true,
1446
+ "should_training_stop": false
1447
+ },
1448
+ "attributes": {}
1449
+ }
1450
+ },
1451
+ "total_flos": 963613493821440.0,
1452
+ "train_batch_size": 1,
1453
+ "trial_name": null,
1454
+ "trial_params": null
1455
+ }
200M/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cda2ce1a3d48dbaae97584542b63fb1b2940959f4e24db1a7c83985cbc8f2f9d
3
+ size 7313