Hugging Face's logo

ugaoo
/
model_32aa9ae6

PEFT
Safetensors
qwen2
Generated from Trainer
4-bit precision
bitsandbytes
Model card Files
xet
 Community
1
ugaoo commited on
Commit
36f0be0
·
verified ·
1 Parent(s): a681330

Upload folder using huggingface_hub

Browse files
Files changed (17) hide show
  1. checkpoint-468/README.md +202 -0
  2. checkpoint-468/adapter_config.json +37 -0
  3. checkpoint-468/adapter_model.safetensors +3 -0
  4. checkpoint-468/added_tokens.json +24 -0
  5. checkpoint-468/global_step466/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
  6. checkpoint-468/global_step466/mp_rank_00_model_states.pt +3 -0
  7. checkpoint-468/latest +1 -0
  8. checkpoint-468/merges.txt +0 -0
  9. checkpoint-468/rng_state.pth +3 -0
  10. checkpoint-468/scheduler.pt +3 -0
  11. checkpoint-468/special_tokens_map.json +31 -0
  12. checkpoint-468/tokenizer.json +3 -0
  13. checkpoint-468/tokenizer_config.json +208 -0
  14. checkpoint-468/trainer_state.json +3309 -0
  15. checkpoint-468/training_args.bin +3 -0
  16. checkpoint-468/vocab.json +0 -0
  17. checkpoint-468/zero_to_fp32.py +604 -0
checkpoint-468/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: Qwen/Qwen2.5-7B-Instruct
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.14.0
checkpoint-468/adapter_config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
5
+ "bias": "none",
6
+ "eva_config": null,
7
+ "exclude_modules": null,
8
+ "fan_in_fan_out": null,
9
+ "inference_mode": true,
10
+ "init_lora_weights": true,
11
+ "layer_replication": null,
12
+ "layers_pattern": null,
13
+ "layers_to_transform": null,
14
+ "loftq_config": {},
15
+ "lora_alpha": 512,
16
+ "lora_bias": false,
17
+ "lora_dropout": 0.05,
18
+ "megatron_config": null,
19
+ "megatron_core": "megatron.core",
20
+ "modules_to_save": null,
21
+ "peft_type": "LORA",
22
+ "r": 256,
23
+ "rank_pattern": {},
24
+ "revision": null,
25
+ "target_modules": [
26
+ "o_proj",
27
+ "v_proj",
28
+ "up_proj",
29
+ "gate_proj",
30
+ "q_proj",
31
+ "k_proj",
32
+ "down_proj"
33
+ ],
34
+ "task_type": "CAUSAL_LM",
35
+ "use_dora": false,
36
+ "use_rslora": false
37
+ }
checkpoint-468/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e9ab88638c8fc4705ac56f34ad42ab86200543b5d5ba0512e832cc2db6c0991
3
+ size 1291899552
checkpoint-468/added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
checkpoint-468/global_step466/bf16_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 sha256:6e770ad451c28f0d134c7434d88482506a57610dbef65138c2dd8fd51d52d5bd
3
+ size 7751120668
checkpoint-468/global_step466/mp_rank_00_model_states.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46e37973c4e235ef28e44acbea22fb3aa4366d800a13e2aa3b8f1e4275fb7fc1
3
+ size 1396311192
checkpoint-468/latest ADDED
@@ -0,0 +1 @@
 
 
1
+ global_step466
checkpoint-468/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-468/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e97952ccfba5983629784a9ebfbcff0a945dacced59660f91df8cd12f1d1af3a
3
+ size 14244
checkpoint-468/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8eddc51b3224c50f48b4a165dd80eef5f713061a1fdddb56c6fa71abf8913466
3
+ size 1064
checkpoint-468/special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
checkpoint-468/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
3
+ size 11421896
checkpoint-468/tokenizer_config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0]['role'] == 'system' %}\n {{- messages[0]['content'] }}\n {%- else %}\n {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n {%- endif %}\n {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0]['role'] == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n {%- else %}\n {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {{- '<|im_start|>' + message.role }}\n {%- if message.content %}\n {{- '\\n' + message.content }}\n {%- endif %}\n {%- for tool_call in message.tool_calls %}\n {%- if tool_call.function is defined %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '\\n<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {{- tool_call.arguments | tojson }}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
199
+ "clean_up_tokenization_spaces": false,
200
+ "eos_token": "<|im_end|>",
201
+ "errors": "replace",
202
+ "extra_special_tokens": {},
203
+ "model_max_length": 131072,
204
+ "pad_token": "<|endoftext|>",
205
+ "split_special_tokens": false,
206
+ "tokenizer_class": "Qwen2Tokenizer",
207
+ "unk_token": null
208
+ }
checkpoint-468/trainer_state.json ADDED
@@ -0,0 +1,3309 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.971590909090909,
5
+ "eval_steps": 500,
6
+ "global_step": 468,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.008522727272727272,
13
+ "grad_norm": 35.49958801269531,
14
+ "learning_rate": 5.0000000000000004e-08,
15
+ "loss": 4.6143,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.017045454545454544,
20
+ "grad_norm": 35.543556213378906,
21
+ "learning_rate": 1.0000000000000001e-07,
22
+ "loss": 4.6719,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.02556818181818182,
27
+ "grad_norm": 34.12852096557617,
28
+ "learning_rate": 1.5000000000000002e-07,
29
+ "loss": 4.5546,
30
+ "step": 3
31
+ },
32
+ {
33
+ "epoch": 0.03409090909090909,
34
+ "grad_norm": 33.610572814941406,
35
+ "learning_rate": 2.0000000000000002e-07,
36
+ "loss": 4.4919,
37
+ "step": 4
38
+ },
39
+ {
40
+ "epoch": 0.04261363636363636,
41
+ "grad_norm": 34.532169342041016,
42
+ "learning_rate": 2.5000000000000004e-07,
43
+ "loss": 4.6095,
44
+ "step": 5
45
+ },
46
+ {
47
+ "epoch": 0.05113636363636364,
48
+ "grad_norm": 34.2357063293457,
49
+ "learning_rate": 3.0000000000000004e-07,
50
+ "loss": 4.585,
51
+ "step": 6
52
+ },
53
+ {
54
+ "epoch": 0.05965909090909091,
55
+ "grad_norm": 34.94609832763672,
56
+ "learning_rate": 3.5000000000000004e-07,
57
+ "loss": 4.5911,
58
+ "step": 7
59
+ },
60
+ {
61
+ "epoch": 0.06818181818181818,
62
+ "grad_norm": 34.79508590698242,
63
+ "learning_rate": 4.0000000000000003e-07,
64
+ "loss": 4.6294,
65
+ "step": 8
66
+ },
67
+ {
68
+ "epoch": 0.07670454545454546,
69
+ "grad_norm": 35.18478775024414,
70
+ "learning_rate": 4.5000000000000003e-07,
71
+ "loss": 4.6568,
72
+ "step": 9
73
+ },
74
+ {
75
+ "epoch": 0.08522727272727272,
76
+ "grad_norm": 33.75633239746094,
77
+ "learning_rate": 5.000000000000001e-07,
78
+ "loss": 4.4504,
79
+ "step": 10
80
+ },
81
+ {
82
+ "epoch": 0.09375,
83
+ "grad_norm": 34.20966339111328,
84
+ "learning_rate": 5.5e-07,
85
+ "loss": 4.5005,
86
+ "step": 11
87
+ },
88
+ {
89
+ "epoch": 0.10227272727272728,
90
+ "grad_norm": 33.20008087158203,
91
+ "learning_rate": 6.000000000000001e-07,
92
+ "loss": 4.4297,
93
+ "step": 12
94
+ },
95
+ {
96
+ "epoch": 0.11079545454545454,
97
+ "grad_norm": 33.53578567504883,
98
+ "learning_rate": 6.5e-07,
99
+ "loss": 4.4136,
100
+ "step": 13
101
+ },
102
+ {
103
+ "epoch": 0.11931818181818182,
104
+ "grad_norm": 31.951068878173828,
105
+ "learning_rate": 7.000000000000001e-07,
106
+ "loss": 4.3065,
107
+ "step": 14
108
+ },
109
+ {
110
+ "epoch": 0.1278409090909091,
111
+ "grad_norm": 30.890714645385742,
112
+ "learning_rate": 7.5e-07,
113
+ "loss": 4.2433,
114
+ "step": 15
115
+ },
116
+ {
117
+ "epoch": 0.13636363636363635,
118
+ "grad_norm": 29.448890686035156,
119
+ "learning_rate": 8.000000000000001e-07,
120
+ "loss": 4.0981,
121
+ "step": 16
122
+ },
123
+ {
124
+ "epoch": 0.14488636363636365,
125
+ "grad_norm": 26.769498825073242,
126
+ "learning_rate": 8.500000000000001e-07,
127
+ "loss": 3.9818,
128
+ "step": 17
129
+ },
130
+ {
131
+ "epoch": 0.1534090909090909,
132
+ "grad_norm": 25.418458938598633,
133
+ "learning_rate": 9.000000000000001e-07,
134
+ "loss": 3.8568,
135
+ "step": 18
136
+ },
137
+ {
138
+ "epoch": 0.16193181818181818,
139
+ "grad_norm": 24.099462509155273,
140
+ "learning_rate": 9.500000000000001e-07,
141
+ "loss": 3.7139,
142
+ "step": 19
143
+ },
144
+ {
145
+ "epoch": 0.17045454545454544,
146
+ "grad_norm": 22.487621307373047,
147
+ "learning_rate": 1.0000000000000002e-06,
148
+ "loss": 3.6967,
149
+ "step": 20
150
+ },
151
+ {
152
+ "epoch": 0.17897727272727273,
153
+ "grad_norm": 20.72856330871582,
154
+ "learning_rate": 1.0500000000000001e-06,
155
+ "loss": 3.5031,
156
+ "step": 21
157
+ },
158
+ {
159
+ "epoch": 0.1875,
160
+ "grad_norm": 19.552040100097656,
161
+ "learning_rate": 1.1e-06,
162
+ "loss": 3.4201,
163
+ "step": 22
164
+ },
165
+ {
166
+ "epoch": 0.19602272727272727,
167
+ "grad_norm": 18.033971786499023,
168
+ "learning_rate": 1.1500000000000002e-06,
169
+ "loss": 3.1842,
170
+ "step": 23
171
+ },
172
+ {
173
+ "epoch": 0.20454545454545456,
174
+ "grad_norm": 18.865802764892578,
175
+ "learning_rate": 1.2000000000000002e-06,
176
+ "loss": 3.1982,
177
+ "step": 24
178
+ },
179
+ {
180
+ "epoch": 0.21306818181818182,
181
+ "grad_norm": 18.849502563476562,
182
+ "learning_rate": 1.25e-06,
183
+ "loss": 2.9954,
184
+ "step": 25
185
+ },
186
+ {
187
+ "epoch": 0.2215909090909091,
188
+ "grad_norm": 19.714330673217773,
189
+ "learning_rate": 1.3e-06,
190
+ "loss": 2.8763,
191
+ "step": 26
192
+ },
193
+ {
194
+ "epoch": 0.23011363636363635,
195
+ "grad_norm": 20.26412010192871,
196
+ "learning_rate": 1.3500000000000002e-06,
197
+ "loss": 2.7259,
198
+ "step": 27
199
+ },
200
+ {
201
+ "epoch": 0.23863636363636365,
202
+ "grad_norm": 19.212318420410156,
203
+ "learning_rate": 1.4000000000000001e-06,
204
+ "loss": 2.6099,
205
+ "step": 28
206
+ },
207
+ {
208
+ "epoch": 0.2471590909090909,
209
+ "grad_norm": 16.80523681640625,
210
+ "learning_rate": 1.45e-06,
211
+ "loss": 2.4482,
212
+ "step": 29
213
+ },
214
+ {
215
+ "epoch": 0.2556818181818182,
216
+ "grad_norm": 14.624052047729492,
217
+ "learning_rate": 1.5e-06,
218
+ "loss": 2.2364,
219
+ "step": 30
220
+ },
221
+ {
222
+ "epoch": 0.26420454545454547,
223
+ "grad_norm": 14.759950637817383,
224
+ "learning_rate": 1.5500000000000002e-06,
225
+ "loss": 2.2159,
226
+ "step": 31
227
+ },
228
+ {
229
+ "epoch": 0.2727272727272727,
230
+ "grad_norm": 14.904441833496094,
231
+ "learning_rate": 1.6000000000000001e-06,
232
+ "loss": 2.0206,
233
+ "step": 32
234
+ },
235
+ {
236
+ "epoch": 0.28125,
237
+ "grad_norm": 15.03490161895752,
238
+ "learning_rate": 1.6500000000000003e-06,
239
+ "loss": 1.8725,
240
+ "step": 33
241
+ },
242
+ {
243
+ "epoch": 0.2897727272727273,
244
+ "grad_norm": 15.70709228515625,
245
+ "learning_rate": 1.7000000000000002e-06,
246
+ "loss": 1.8046,
247
+ "step": 34
248
+ },
249
+ {
250
+ "epoch": 0.29829545454545453,
251
+ "grad_norm": 15.637526512145996,
252
+ "learning_rate": 1.75e-06,
253
+ "loss": 1.6198,
254
+ "step": 35
255
+ },
256
+ {
257
+ "epoch": 0.3068181818181818,
258
+ "grad_norm": 14.345853805541992,
259
+ "learning_rate": 1.8000000000000001e-06,
260
+ "loss": 1.4247,
261
+ "step": 36
262
+ },
263
+ {
264
+ "epoch": 0.3153409090909091,
265
+ "grad_norm": 14.281502723693848,
266
+ "learning_rate": 1.85e-06,
267
+ "loss": 1.2859,
268
+ "step": 37
269
+ },
270
+ {
271
+ "epoch": 0.32386363636363635,
272
+ "grad_norm": 13.567434310913086,
273
+ "learning_rate": 1.9000000000000002e-06,
274
+ "loss": 1.1452,
275
+ "step": 38
276
+ },
277
+ {
278
+ "epoch": 0.33238636363636365,
279
+ "grad_norm": 13.128920555114746,
280
+ "learning_rate": 1.9500000000000004e-06,
281
+ "loss": 1.0036,
282
+ "step": 39
283
+ },
284
+ {
285
+ "epoch": 0.3409090909090909,
286
+ "grad_norm": 12.954020500183105,
287
+ "learning_rate": 2.0000000000000003e-06,
288
+ "loss": 0.851,
289
+ "step": 40
290
+ },
291
+ {
292
+ "epoch": 0.3494318181818182,
293
+ "grad_norm": 12.533946990966797,
294
+ "learning_rate": 2.05e-06,
295
+ "loss": 0.715,
296
+ "step": 41
297
+ },
298
+ {
299
+ "epoch": 0.35795454545454547,
300
+ "grad_norm": 11.564764022827148,
301
+ "learning_rate": 2.1000000000000002e-06,
302
+ "loss": 0.5987,
303
+ "step": 42
304
+ },
305
+ {
306
+ "epoch": 0.3664772727272727,
307
+ "grad_norm": 10.383822441101074,
308
+ "learning_rate": 2.15e-06,
309
+ "loss": 0.4834,
310
+ "step": 43
311
+ },
312
+ {
313
+ "epoch": 0.375,
314
+ "grad_norm": 8.901700973510742,
315
+ "learning_rate": 2.2e-06,
316
+ "loss": 0.3808,
317
+ "step": 44
318
+ },
319
+ {
320
+ "epoch": 0.3835227272727273,
321
+ "grad_norm": 7.737320423126221,
322
+ "learning_rate": 2.25e-06,
323
+ "loss": 0.2975,
324
+ "step": 45
325
+ },
326
+ {
327
+ "epoch": 0.39204545454545453,
328
+ "grad_norm": 5.334733963012695,
329
+ "learning_rate": 2.3000000000000004e-06,
330
+ "loss": 0.2261,
331
+ "step": 46
332
+ },
333
+ {
334
+ "epoch": 0.4005681818181818,
335
+ "grad_norm": 3.3499741554260254,
336
+ "learning_rate": 2.35e-06,
337
+ "loss": 0.1854,
338
+ "step": 47
339
+ },
340
+ {
341
+ "epoch": 0.4090909090909091,
342
+ "grad_norm": 2.4037742614746094,
343
+ "learning_rate": 2.4000000000000003e-06,
344
+ "loss": 0.1532,
345
+ "step": 48
346
+ },
347
+ {
348
+ "epoch": 0.41761363636363635,
349
+ "grad_norm": 1.8914185762405396,
350
+ "learning_rate": 2.4500000000000003e-06,
351
+ "loss": 0.1372,
352
+ "step": 49
353
+ },
354
+ {
355
+ "epoch": 0.42613636363636365,
356
+ "grad_norm": 1.7124507427215576,
357
+ "learning_rate": 2.5e-06,
358
+ "loss": 0.136,
359
+ "step": 50
360
+ },
361
+ {
362
+ "epoch": 0.4346590909090909,
363
+ "grad_norm": 1.242527961730957,
364
+ "learning_rate": 2.55e-06,
365
+ "loss": 0.1238,
366
+ "step": 51
367
+ },
368
+ {
369
+ "epoch": 0.4431818181818182,
370
+ "grad_norm": 0.9835780262947083,
371
+ "learning_rate": 2.6e-06,
372
+ "loss": 0.1192,
373
+ "step": 52
374
+ },
375
+ {
376
+ "epoch": 0.45170454545454547,
377
+ "grad_norm": 1.0163497924804688,
378
+ "learning_rate": 2.6500000000000005e-06,
379
+ "loss": 0.1175,
380
+ "step": 53
381
+ },
382
+ {
383
+ "epoch": 0.4602272727272727,
384
+ "grad_norm": 0.8837094306945801,
385
+ "learning_rate": 2.7000000000000004e-06,
386
+ "loss": 0.1104,
387
+ "step": 54
388
+ },
389
+ {
390
+ "epoch": 0.46875,
391
+ "grad_norm": 0.6102741956710815,
392
+ "learning_rate": 2.7500000000000004e-06,
393
+ "loss": 0.0986,
394
+ "step": 55
395
+ },
396
+ {
397
+ "epoch": 0.4772727272727273,
398
+ "grad_norm": 0.85715252161026,
399
+ "learning_rate": 2.8000000000000003e-06,
400
+ "loss": 0.1083,
401
+ "step": 56
402
+ },
403
+ {
404
+ "epoch": 0.48579545454545453,
405
+ "grad_norm": 0.9692059755325317,
406
+ "learning_rate": 2.85e-06,
407
+ "loss": 0.0994,
408
+ "step": 57
409
+ },
410
+ {
411
+ "epoch": 0.4943181818181818,
412
+ "grad_norm": 0.5620752573013306,
413
+ "learning_rate": 2.9e-06,
414
+ "loss": 0.0909,
415
+ "step": 58
416
+ },
417
+ {
418
+ "epoch": 0.5028409090909091,
419
+ "grad_norm": 0.5377550721168518,
420
+ "learning_rate": 2.95e-06,
421
+ "loss": 0.087,
422
+ "step": 59
423
+ },
424
+ {
425
+ "epoch": 0.5113636363636364,
426
+ "grad_norm": 0.7260486483573914,
427
+ "learning_rate": 3e-06,
428
+ "loss": 0.0949,
429
+ "step": 60
430
+ },
431
+ {
432
+ "epoch": 0.5198863636363636,
433
+ "grad_norm": 0.5636699199676514,
434
+ "learning_rate": 3.05e-06,
435
+ "loss": 0.0884,
436
+ "step": 61
437
+ },
438
+ {
439
+ "epoch": 0.5284090909090909,
440
+ "grad_norm": 0.3729614019393921,
441
+ "learning_rate": 3.1000000000000004e-06,
442
+ "loss": 0.0771,
443
+ "step": 62
444
+ },
445
+ {
446
+ "epoch": 0.5369318181818182,
447
+ "grad_norm": 0.5472120046615601,
448
+ "learning_rate": 3.1500000000000003e-06,
449
+ "loss": 0.0887,
450
+ "step": 63
451
+ },
452
+ {
453
+ "epoch": 0.5454545454545454,
454
+ "grad_norm": 0.4915490448474884,
455
+ "learning_rate": 3.2000000000000003e-06,
456
+ "loss": 0.0826,
457
+ "step": 64
458
+ },
459
+ {
460
+ "epoch": 0.5539772727272727,
461
+ "grad_norm": 0.4123076796531677,
462
+ "learning_rate": 3.2500000000000002e-06,
463
+ "loss": 0.0834,
464
+ "step": 65
465
+ },
466
+ {
467
+ "epoch": 0.5625,
468
+ "grad_norm": 0.32767826318740845,
469
+ "learning_rate": 3.3000000000000006e-06,
470
+ "loss": 0.0765,
471
+ "step": 66
472
+ },
473
+ {
474
+ "epoch": 0.5710227272727273,
475
+ "grad_norm": 0.34970155358314514,
476
+ "learning_rate": 3.3500000000000005e-06,
477
+ "loss": 0.0832,
478
+ "step": 67
479
+ },
480
+ {
481
+ "epoch": 0.5795454545454546,
482
+ "grad_norm": 0.38679587841033936,
483
+ "learning_rate": 3.4000000000000005e-06,
484
+ "loss": 0.0766,
485
+ "step": 68
486
+ },
487
+ {
488
+ "epoch": 0.5880681818181818,
489
+ "grad_norm": 0.5464550852775574,
490
+ "learning_rate": 3.45e-06,
491
+ "loss": 0.0748,
492
+ "step": 69
493
+ },
494
+ {
495
+ "epoch": 0.5965909090909091,
496
+ "grad_norm": 0.3545376658439636,
497
+ "learning_rate": 3.5e-06,
498
+ "loss": 0.0759,
499
+ "step": 70
500
+ },
501
+ {
502
+ "epoch": 0.6051136363636364,
503
+ "grad_norm": 0.3532780110836029,
504
+ "learning_rate": 3.5500000000000003e-06,
505
+ "loss": 0.0728,
506
+ "step": 71
507
+ },
508
+ {
509
+ "epoch": 0.6136363636363636,
510
+ "grad_norm": 0.5024192929267883,
511
+ "learning_rate": 3.6000000000000003e-06,
512
+ "loss": 0.081,
513
+ "step": 72
514
+ },
515
+ {
516
+ "epoch": 0.6221590909090909,
517
+ "grad_norm": 0.3845844268798828,
518
+ "learning_rate": 3.65e-06,
519
+ "loss": 0.0689,
520
+ "step": 73
521
+ },
522
+ {
523
+ "epoch": 0.6306818181818182,
524
+ "grad_norm": 0.46010249853134155,
525
+ "learning_rate": 3.7e-06,
526
+ "loss": 0.0757,
527
+ "step": 74
528
+ },
529
+ {
530
+ "epoch": 0.6392045454545454,
531
+ "grad_norm": 0.36048972606658936,
532
+ "learning_rate": 3.7500000000000005e-06,
533
+ "loss": 0.0758,
534
+ "step": 75
535
+ },
536
+ {
537
+ "epoch": 0.6477272727272727,
538
+ "grad_norm": 0.2978876829147339,
539
+ "learning_rate": 3.8000000000000005e-06,
540
+ "loss": 0.0747,
541
+ "step": 76
542
+ },
543
+ {
544
+ "epoch": 0.65625,
545
+ "grad_norm": 0.30899888277053833,
546
+ "learning_rate": 3.85e-06,
547
+ "loss": 0.0788,
548
+ "step": 77
549
+ },
550
+ {
551
+ "epoch": 0.6647727272727273,
552
+ "grad_norm": 0.45000651478767395,
553
+ "learning_rate": 3.900000000000001e-06,
554
+ "loss": 0.0672,
555
+ "step": 78
556
+ },
557
+ {
558
+ "epoch": 0.6732954545454546,
559
+ "grad_norm": 0.2621772587299347,
560
+ "learning_rate": 3.95e-06,
561
+ "loss": 0.0719,
562
+ "step": 79
563
+ },
564
+ {
565
+ "epoch": 0.6818181818181818,
566
+ "grad_norm": 0.28618085384368896,
567
+ "learning_rate": 4.000000000000001e-06,
568
+ "loss": 0.0744,
569
+ "step": 80
570
+ },
571
+ {
572
+ "epoch": 0.6903409090909091,
573
+ "grad_norm": 0.40642479062080383,
574
+ "learning_rate": 4.05e-06,
575
+ "loss": 0.0743,
576
+ "step": 81
577
+ },
578
+ {
579
+ "epoch": 0.6988636363636364,
580
+ "grad_norm": 0.3514344096183777,
581
+ "learning_rate": 4.1e-06,
582
+ "loss": 0.0708,
583
+ "step": 82
584
+ },
585
+ {
586
+ "epoch": 0.7073863636363636,
587
+ "grad_norm": 0.426798939704895,
588
+ "learning_rate": 4.15e-06,
589
+ "loss": 0.073,
590
+ "step": 83
591
+ },
592
+ {
593
+ "epoch": 0.7159090909090909,
594
+ "grad_norm": 0.29413217306137085,
595
+ "learning_rate": 4.2000000000000004e-06,
596
+ "loss": 0.0732,
597
+ "step": 84
598
+ },
599
+ {
600
+ "epoch": 0.7244318181818182,
601
+ "grad_norm": 0.37668099999427795,
602
+ "learning_rate": 4.25e-06,
603
+ "loss": 0.0664,
604
+ "step": 85
605
+ },
606
+ {
607
+ "epoch": 0.7329545454545454,
608
+ "grad_norm": 0.3696061372756958,
609
+ "learning_rate": 4.3e-06,
610
+ "loss": 0.0649,
611
+ "step": 86
612
+ },
613
+ {
614
+ "epoch": 0.7414772727272727,
615
+ "grad_norm": 0.311988890171051,
616
+ "learning_rate": 4.350000000000001e-06,
617
+ "loss": 0.0689,
618
+ "step": 87
619
+ },
620
+ {
621
+ "epoch": 0.75,
622
+ "grad_norm": 0.3180747628211975,
623
+ "learning_rate": 4.4e-06,
624
+ "loss": 0.0746,
625
+ "step": 88
626
+ },
627
+ {
628
+ "epoch": 0.7585227272727273,
629
+ "grad_norm": 0.46045729517936707,
630
+ "learning_rate": 4.450000000000001e-06,
631
+ "loss": 0.0763,
632
+ "step": 89
633
+ },
634
+ {
635
+ "epoch": 0.7670454545454546,
636
+ "grad_norm": 0.3566094934940338,
637
+ "learning_rate": 4.5e-06,
638
+ "loss": 0.0676,
639
+ "step": 90
640
+ },
641
+ {
642
+ "epoch": 0.7755681818181818,
643
+ "grad_norm": 0.35632985830307007,
644
+ "learning_rate": 4.5500000000000005e-06,
645
+ "loss": 0.0677,
646
+ "step": 91
647
+ },
648
+ {
649
+ "epoch": 0.7840909090909091,
650
+ "grad_norm": 0.27962526679039,
651
+ "learning_rate": 4.600000000000001e-06,
652
+ "loss": 0.0689,
653
+ "step": 92
654
+ },
655
+ {
656
+ "epoch": 0.7926136363636364,
657
+ "grad_norm": 0.4532427191734314,
658
+ "learning_rate": 4.65e-06,
659
+ "loss": 0.0724,
660
+ "step": 93
661
+ },
662
+ {
663
+ "epoch": 0.8011363636363636,
664
+ "grad_norm": 0.363337904214859,
665
+ "learning_rate": 4.7e-06,
666
+ "loss": 0.0708,
667
+ "step": 94
668
+ },
669
+ {
670
+ "epoch": 0.8096590909090909,
671
+ "grad_norm": 0.3065521717071533,
672
+ "learning_rate": 4.75e-06,
673
+ "loss": 0.0713,
674
+ "step": 95
675
+ },
676
+ {
677
+ "epoch": 0.8181818181818182,
678
+ "grad_norm": 0.24705548584461212,
679
+ "learning_rate": 4.800000000000001e-06,
680
+ "loss": 0.0683,
681
+ "step": 96
682
+ },
683
+ {
684
+ "epoch": 0.8267045454545454,
685
+ "grad_norm": 0.5038250684738159,
686
+ "learning_rate": 4.85e-06,
687
+ "loss": 0.0738,
688
+ "step": 97
689
+ },
690
+ {
691
+ "epoch": 0.8352272727272727,
692
+ "grad_norm": 0.2972690761089325,
693
+ "learning_rate": 4.9000000000000005e-06,
694
+ "loss": 0.0684,
695
+ "step": 98
696
+ },
697
+ {
698
+ "epoch": 0.84375,
699
+ "grad_norm": 0.40811270475387573,
700
+ "learning_rate": 4.95e-06,
701
+ "loss": 0.066,
702
+ "step": 99
703
+ },
704
+ {
705
+ "epoch": 0.8522727272727273,
706
+ "grad_norm": 0.3925577998161316,
707
+ "learning_rate": 5e-06,
708
+ "loss": 0.0692,
709
+ "step": 100
710
+ },
711
+ {
712
+ "epoch": 0.8607954545454546,
713
+ "grad_norm": 0.32043716311454773,
714
+ "learning_rate": 4.999965957943338e-06,
715
+ "loss": 0.0652,
716
+ "step": 101
717
+ },
718
+ {
719
+ "epoch": 0.8693181818181818,
720
+ "grad_norm": 0.26025497913360596,
721
+ "learning_rate": 4.999863832700438e-06,
722
+ "loss": 0.0634,
723
+ "step": 102
724
+ },
725
+ {
726
+ "epoch": 0.8778409090909091,
727
+ "grad_norm": 0.5605457425117493,
728
+ "learning_rate": 4.999693627052545e-06,
729
+ "loss": 0.0686,
730
+ "step": 103
731
+ },
732
+ {
733
+ "epoch": 0.8863636363636364,
734
+ "grad_norm": 0.48003220558166504,
735
+ "learning_rate": 4.9994553456349785e-06,
736
+ "loss": 0.0707,
737
+ "step": 104
738
+ },
739
+ {
740
+ "epoch": 0.8948863636363636,
741
+ "grad_norm": 0.22346197068691254,
742
+ "learning_rate": 4.99914899493701e-06,
743
+ "loss": 0.067,
744
+ "step": 105
745
+ },
746
+ {
747
+ "epoch": 0.9034090909090909,
748
+ "grad_norm": 0.45966312289237976,
749
+ "learning_rate": 4.998774583301685e-06,
750
+ "loss": 0.0663,
751
+ "step": 106
752
+ },
753
+ {
754
+ "epoch": 0.9119318181818182,
755
+ "grad_norm": 0.38695356249809265,
756
+ "learning_rate": 4.998332120925598e-06,
757
+ "loss": 0.0714,
758
+ "step": 107
759
+ },
760
+ {
761
+ "epoch": 0.9204545454545454,
762
+ "grad_norm": 0.35708868503570557,
763
+ "learning_rate": 4.997821619858614e-06,
764
+ "loss": 0.0632,
765
+ "step": 108
766
+ },
767
+ {
768
+ "epoch": 0.9289772727272727,
769
+ "grad_norm": 0.2699528932571411,
770
+ "learning_rate": 4.9972430940035355e-06,
771
+ "loss": 0.0649,
772
+ "step": 109
773
+ },
774
+ {
775
+ "epoch": 0.9375,
776
+ "grad_norm": 0.5043486952781677,
777
+ "learning_rate": 4.9965965591157314e-06,
778
+ "loss": 0.0715,
779
+ "step": 110
780
+ },
781
+ {
782
+ "epoch": 0.9460227272727273,
783
+ "grad_norm": 0.4046338200569153,
784
+ "learning_rate": 4.995882032802703e-06,
785
+ "loss": 0.0657,
786
+ "step": 111
787
+ },
788
+ {
789
+ "epoch": 0.9545454545454546,
790
+ "grad_norm": 0.3543775677680969,
791
+ "learning_rate": 4.995099534523608e-06,
792
+ "loss": 0.0698,
793
+ "step": 112
794
+ },
795
+ {
796
+ "epoch": 0.9630681818181818,
797
+ "grad_norm": 0.3295219838619232,
798
+ "learning_rate": 4.994249085588725e-06,
799
+ "loss": 0.0719,
800
+ "step": 113
801
+ },
802
+ {
803
+ "epoch": 0.9715909090909091,
804
+ "grad_norm": 0.3691340982913971,
805
+ "learning_rate": 4.993330709158879e-06,
806
+ "loss": 0.0689,
807
+ "step": 114
808
+ },
809
+ {
810
+ "epoch": 0.9801136363636364,
811
+ "grad_norm": 0.4219862222671509,
812
+ "learning_rate": 4.9923444302448095e-06,
813
+ "loss": 0.0644,
814
+ "step": 115
815
+ },
816
+ {
817
+ "epoch": 0.9886363636363636,
818
+ "grad_norm": 0.33437392115592957,
819
+ "learning_rate": 4.991290275706486e-06,
820
+ "loss": 0.074,
821
+ "step": 116
822
+ },
823
+ {
824
+ "epoch": 0.9971590909090909,
825
+ "grad_norm": 0.22970478236675262,
826
+ "learning_rate": 4.990168274252379e-06,
827
+ "loss": 0.0639,
828
+ "step": 117
829
+ },
830
+ {
831
+ "epoch": 1.0,
832
+ "grad_norm": 0.22970478236675262,
833
+ "learning_rate": 4.988978456438678e-06,
834
+ "loss": 0.0718,
835
+ "step": 118
836
+ },
837
+ {
838
+ "epoch": 1.0085227272727273,
839
+ "grad_norm": 0.7316553592681885,
840
+ "learning_rate": 4.98772085466846e-06,
841
+ "loss": 0.062,
842
+ "step": 119
843
+ },
844
+ {
845
+ "epoch": 1.0170454545454546,
846
+ "grad_norm": 0.27226850390434265,
847
+ "learning_rate": 4.986395503190805e-06,
848
+ "loss": 0.0583,
849
+ "step": 120
850
+ },
851
+ {
852
+ "epoch": 1.0255681818181819,
853
+ "grad_norm": 0.20066767930984497,
854
+ "learning_rate": 4.9850024380998655e-06,
855
+ "loss": 0.0624,
856
+ "step": 121
857
+ },
858
+ {
859
+ "epoch": 1.0340909090909092,
860
+ "grad_norm": 0.26274779438972473,
861
+ "learning_rate": 4.9835416973338815e-06,
862
+ "loss": 0.0641,
863
+ "step": 122
864
+ },
865
+ {
866
+ "epoch": 1.0426136363636365,
867
+ "grad_norm": 0.3893499970436096,
868
+ "learning_rate": 4.982013320674149e-06,
869
+ "loss": 0.066,
870
+ "step": 123
871
+ },
872
+ {
873
+ "epoch": 1.0511363636363635,
874
+ "grad_norm": 0.3454163670539856,
875
+ "learning_rate": 4.980417349743936e-06,
876
+ "loss": 0.061,
877
+ "step": 124
878
+ },
879
+ {
880
+ "epoch": 1.0596590909090908,
881
+ "grad_norm": 0.26956701278686523,
882
+ "learning_rate": 4.97875382800735e-06,
883
+ "loss": 0.0641,
884
+ "step": 125
885
+ },
886
+ {
887
+ "epoch": 1.0681818181818181,
888
+ "grad_norm": 0.4147774279117584,
889
+ "learning_rate": 4.97702280076815e-06,
890
+ "loss": 0.0599,
891
+ "step": 126
892
+ },
893
+ {
894
+ "epoch": 1.0767045454545454,
895
+ "grad_norm": 0.2839421033859253,
896
+ "learning_rate": 4.9752243151685185e-06,
897
+ "loss": 0.0669,
898
+ "step": 127
899
+ },
900
+ {
901
+ "epoch": 1.0852272727272727,
902
+ "grad_norm": 0.41760239005088806,
903
+ "learning_rate": 4.973358420187776e-06,
904
+ "loss": 0.0661,
905
+ "step": 128
906
+ },
907
+ {
908
+ "epoch": 1.09375,
909
+ "grad_norm": 0.29833659529685974,
910
+ "learning_rate": 4.9714251666410426e-06,
911
+ "loss": 0.0672,
912
+ "step": 129
913
+ },
914
+ {
915
+ "epoch": 1.1022727272727273,
916
+ "grad_norm": 0.2806449830532074,
917
+ "learning_rate": 4.969424607177861e-06,
918
+ "loss": 0.0589,
919
+ "step": 130
920
+ },
921
+ {
922
+ "epoch": 1.1107954545454546,
923
+ "grad_norm": 0.33876338601112366,
924
+ "learning_rate": 4.967356796280756e-06,
925
+ "loss": 0.0647,
926
+ "step": 131
927
+ },
928
+ {
929
+ "epoch": 1.1193181818181819,
930
+ "grad_norm": 0.2331860363483429,
931
+ "learning_rate": 4.96522179026376e-06,
932
+ "loss": 0.0633,
933
+ "step": 132
934
+ },
935
+ {
936
+ "epoch": 1.1278409090909092,
937
+ "grad_norm": 0.22244933247566223,
938
+ "learning_rate": 4.963019647270866e-06,
939
+ "loss": 0.0618,
940
+ "step": 133
941
+ },
942
+ {
943
+ "epoch": 1.1363636363636362,
944
+ "grad_norm": 0.24649232625961304,
945
+ "learning_rate": 4.960750427274458e-06,
946
+ "loss": 0.0565,
947
+ "step": 134
948
+ },
949
+ {
950
+ "epoch": 1.1448863636363638,
951
+ "grad_norm": 0.23454727232456207,
952
+ "learning_rate": 4.958414192073665e-06,
953
+ "loss": 0.0607,
954
+ "step": 135
955
+ },
956
+ {
957
+ "epoch": 1.1534090909090908,
958
+ "grad_norm": 0.28441867232322693,
959
+ "learning_rate": 4.956011005292693e-06,
960
+ "loss": 0.0548,
961
+ "step": 136
962
+ },
963
+ {
964
+ "epoch": 1.1619318181818181,
965
+ "grad_norm": 0.26571962237358093,
966
+ "learning_rate": 4.9535409323790726e-06,
967
+ "loss": 0.0596,
968
+ "step": 137
969
+ },
970
+ {
971
+ "epoch": 1.1704545454545454,
972
+ "grad_norm": 0.2692795991897583,
973
+ "learning_rate": 4.951004040601898e-06,
974
+ "loss": 0.0572,
975
+ "step": 138
976
+ },
977
+ {
978
+ "epoch": 1.1789772727272727,
979
+ "grad_norm": 0.35209110379219055,
980
+ "learning_rate": 4.9484003990499785e-06,
981
+ "loss": 0.0657,
982
+ "step": 139
983
+ },
984
+ {
985
+ "epoch": 1.1875,
986
+ "grad_norm": 0.2726614773273468,
987
+ "learning_rate": 4.945730078629965e-06,
988
+ "loss": 0.0561,
989
+ "step": 140
990
+ },
991
+ {
992
+ "epoch": 1.1960227272727273,
993
+ "grad_norm": 0.2439083456993103,
994
+ "learning_rate": 4.942993152064415e-06,
995
+ "loss": 0.0594,
996
+ "step": 141
997
+ },
998
+ {
999
+ "epoch": 1.2045454545454546,
1000
+ "grad_norm": 0.23282821476459503,
1001
+ "learning_rate": 4.940189693889819e-06,
1002
+ "loss": 0.0598,
1003
+ "step": 142
1004
+ },
1005
+ {
1006
+ "epoch": 1.2130681818181819,
1007
+ "grad_norm": 0.22507373988628387,
1008
+ "learning_rate": 4.937319780454559e-06,
1009
+ "loss": 0.0632,
1010
+ "step": 143
1011
+ },
1012
+ {
1013
+ "epoch": 1.2215909090909092,
1014
+ "grad_norm": 0.35792839527130127,
1015
+ "learning_rate": 4.934383489916843e-06,
1016
+ "loss": 0.0626,
1017
+ "step": 144
1018
+ },
1019
+ {
1020
+ "epoch": 1.2301136363636362,
1021
+ "grad_norm": 0.3073280453681946,
1022
+ "learning_rate": 4.931380902242564e-06,
1023
+ "loss": 0.0558,
1024
+ "step": 145
1025
+ },
1026
+ {
1027
+ "epoch": 1.2386363636363638,
1028
+ "grad_norm": 0.2743082642555237,
1029
+ "learning_rate": 4.928312099203131e-06,
1030
+ "loss": 0.0638,
1031
+ "step": 146
1032
+ },
1033
+ {
1034
+ "epoch": 1.2471590909090908,
1035
+ "grad_norm": 0.2522992491722107,
1036
+ "learning_rate": 4.925177164373237e-06,
1037
+ "loss": 0.0599,
1038
+ "step": 147
1039
+ },
1040
+ {
1041
+ "epoch": 1.2556818181818181,
1042
+ "grad_norm": 0.25230202078819275,
1043
+ "learning_rate": 4.921976183128585e-06,
1044
+ "loss": 0.0579,
1045
+ "step": 148
1046
+ },
1047
+ {
1048
+ "epoch": 1.2642045454545454,
1049
+ "grad_norm": 0.24467583000659943,
1050
+ "learning_rate": 4.918709242643563e-06,
1051
+ "loss": 0.0607,
1052
+ "step": 149
1053
+ },
1054
+ {
1055
+ "epoch": 1.2727272727272727,
1056
+ "grad_norm": 0.2191130518913269,
1057
+ "learning_rate": 4.915376431888871e-06,
1058
+ "loss": 0.0622,
1059
+ "step": 150
1060
+ },
1061
+ {
1062
+ "epoch": 1.28125,
1063
+ "grad_norm": 0.26959559321403503,
1064
+ "learning_rate": 4.911977841629092e-06,
1065
+ "loss": 0.0568,
1066
+ "step": 151
1067
+ },
1068
+ {
1069
+ "epoch": 1.2897727272727273,
1070
+ "grad_norm": 0.25139617919921875,
1071
+ "learning_rate": 4.908513564420231e-06,
1072
+ "loss": 0.0607,
1073
+ "step": 152
1074
+ },
1075
+ {
1076
+ "epoch": 1.2982954545454546,
1077
+ "grad_norm": 0.27705878019332886,
1078
+ "learning_rate": 4.904983694607183e-06,
1079
+ "loss": 0.0607,
1080
+ "step": 153
1081
+ },
1082
+ {
1083
+ "epoch": 1.3068181818181819,
1084
+ "grad_norm": 0.2860431373119354,
1085
+ "learning_rate": 4.9013883283211705e-06,
1086
+ "loss": 0.0572,
1087
+ "step": 154
1088
+ },
1089
+ {
1090
+ "epoch": 1.3153409090909092,
1091
+ "grad_norm": 0.27946561574935913,
1092
+ "learning_rate": 4.897727563477123e-06,
1093
+ "loss": 0.054,
1094
+ "step": 155
1095
+ },
1096
+ {
1097
+ "epoch": 1.3238636363636362,
1098
+ "grad_norm": 0.25392627716064453,
1099
+ "learning_rate": 4.894001499771015e-06,
1100
+ "loss": 0.0623,
1101
+ "step": 156
1102
+ },
1103
+ {
1104
+ "epoch": 1.3323863636363638,
1105
+ "grad_norm": 0.24209532141685486,
1106
+ "learning_rate": 4.890210238677141e-06,
1107
+ "loss": 0.0592,
1108
+ "step": 157
1109
+ },
1110
+ {
1111
+ "epoch": 1.3409090909090908,
1112
+ "grad_norm": 0.23799730837345123,
1113
+ "learning_rate": 4.886353883445363e-06,
1114
+ "loss": 0.0545,
1115
+ "step": 158
1116
+ },
1117
+ {
1118
+ "epoch": 1.3494318181818181,
1119
+ "grad_norm": 0.20330385863780975,
1120
+ "learning_rate": 4.88243253909829e-06,
1121
+ "loss": 0.0582,
1122
+ "step": 159
1123
+ },
1124
+ {
1125
+ "epoch": 1.3579545454545454,
1126
+ "grad_norm": 0.2390514761209488,
1127
+ "learning_rate": 4.878446312428424e-06,
1128
+ "loss": 0.0558,
1129
+ "step": 160
1130
+ },
1131
+ {
1132
+ "epoch": 1.3664772727272727,
1133
+ "grad_norm": 0.3132871687412262,
1134
+ "learning_rate": 4.874395311995248e-06,
1135
+ "loss": 0.0522,
1136
+ "step": 161
1137
+ },
1138
+ {
1139
+ "epoch": 1.375,
1140
+ "grad_norm": 0.2441999763250351,
1141
+ "learning_rate": 4.870279648122271e-06,
1142
+ "loss": 0.0632,
1143
+ "step": 162
1144
+ },
1145
+ {
1146
+ "epoch": 1.3835227272727273,
1147
+ "grad_norm": 0.3468320369720459,
1148
+ "learning_rate": 4.866099432894023e-06,
1149
+ "loss": 0.0544,
1150
+ "step": 163
1151
+ },
1152
+ {
1153
+ "epoch": 1.3920454545454546,
1154
+ "grad_norm": 0.2637532949447632,
1155
+ "learning_rate": 4.8618547801530045e-06,
1156
+ "loss": 0.0508,
1157
+ "step": 164
1158
+ },
1159
+ {
1160
+ "epoch": 1.4005681818181819,
1161
+ "grad_norm": 0.25950244069099426,
1162
+ "learning_rate": 4.857545805496581e-06,
1163
+ "loss": 0.0535,
1164
+ "step": 165
1165
+ },
1166
+ {
1167
+ "epoch": 1.4090909090909092,
1168
+ "grad_norm": 0.24120447039604187,
1169
+ "learning_rate": 4.853172626273841e-06,
1170
+ "loss": 0.0563,
1171
+ "step": 166
1172
+ },
1173
+ {
1174
+ "epoch": 1.4176136363636362,
1175
+ "grad_norm": 0.2234521359205246,
1176
+ "learning_rate": 4.848735361582396e-06,
1177
+ "loss": 0.0526,
1178
+ "step": 167
1179
+ },
1180
+ {
1181
+ "epoch": 1.4261363636363638,
1182
+ "grad_norm": 0.24657292664051056,
1183
+ "learning_rate": 4.844234132265139e-06,
1184
+ "loss": 0.0596,
1185
+ "step": 168
1186
+ },
1187
+ {
1188
+ "epoch": 1.4346590909090908,
1189
+ "grad_norm": 0.4169978201389313,
1190
+ "learning_rate": 4.839669060906952e-06,
1191
+ "loss": 0.0569,
1192
+ "step": 169
1193
+ },
1194
+ {
1195
+ "epoch": 1.4431818181818181,
1196
+ "grad_norm": 0.24876588582992554,
1197
+ "learning_rate": 4.835040271831371e-06,
1198
+ "loss": 0.0555,
1199
+ "step": 170
1200
+ },
1201
+ {
1202
+ "epoch": 1.4517045454545454,
1203
+ "grad_norm": 0.2905561923980713,
1204
+ "learning_rate": 4.830347891097193e-06,
1205
+ "loss": 0.0518,
1206
+ "step": 171
1207
+ },
1208
+ {
1209
+ "epoch": 1.4602272727272727,
1210
+ "grad_norm": 0.32290127873420715,
1211
+ "learning_rate": 4.8255920464950545e-06,
1212
+ "loss": 0.0544,
1213
+ "step": 172
1214
+ },
1215
+ {
1216
+ "epoch": 1.46875,
1217
+ "grad_norm": 0.3065219223499298,
1218
+ "learning_rate": 4.820772867543939e-06,
1219
+ "loss": 0.0563,
1220
+ "step": 173
1221
+ },
1222
+ {
1223
+ "epoch": 1.4772727272727273,
1224
+ "grad_norm": 0.3081221878528595,
1225
+ "learning_rate": 4.8158904854876555e-06,
1226
+ "loss": 0.0595,
1227
+ "step": 174
1228
+ },
1229
+ {
1230
+ "epoch": 1.4857954545454546,
1231
+ "grad_norm": 0.3171551823616028,
1232
+ "learning_rate": 4.810945033291267e-06,
1233
+ "loss": 0.0509,
1234
+ "step": 175
1235
+ },
1236
+ {
1237
+ "epoch": 1.4943181818181819,
1238
+ "grad_norm": 0.46660706400871277,
1239
+ "learning_rate": 4.805936645637463e-06,
1240
+ "loss": 0.0575,
1241
+ "step": 176
1242
+ },
1243
+ {
1244
+ "epoch": 1.5028409090909092,
1245
+ "grad_norm": 0.3173319399356842,
1246
+ "learning_rate": 4.800865458922899e-06,
1247
+ "loss": 0.0579,
1248
+ "step": 177
1249
+ },
1250
+ {
1251
+ "epoch": 1.5113636363636362,
1252
+ "grad_norm": 0.2553943693637848,
1253
+ "learning_rate": 4.795731611254473e-06,
1254
+ "loss": 0.0602,
1255
+ "step": 178
1256
+ },
1257
+ {
1258
+ "epoch": 1.5198863636363638,
1259
+ "grad_norm": 0.39231568574905396,
1260
+ "learning_rate": 4.790535242445573e-06,
1261
+ "loss": 0.0606,
1262
+ "step": 179
1263
+ },
1264
+ {
1265
+ "epoch": 1.5284090909090908,
1266
+ "grad_norm": 0.2721359133720398,
1267
+ "learning_rate": 4.7852764940122636e-06,
1268
+ "loss": 0.0545,
1269
+ "step": 180
1270
+ },
1271
+ {
1272
+ "epoch": 1.5369318181818183,
1273
+ "grad_norm": 0.2545195519924164,
1274
+ "learning_rate": 4.779955509169435e-06,
1275
+ "loss": 0.0546,
1276
+ "step": 181
1277
+ },
1278
+ {
1279
+ "epoch": 1.5454545454545454,
1280
+ "grad_norm": 0.3155561685562134,
1281
+ "learning_rate": 4.7745724328269e-06,
1282
+ "loss": 0.0543,
1283
+ "step": 182
1284
+ },
1285
+ {
1286
+ "epoch": 1.5539772727272727,
1287
+ "grad_norm": 0.24933341145515442,
1288
+ "learning_rate": 4.769127411585452e-06,
1289
+ "loss": 0.0571,
1290
+ "step": 183
1291
+ },
1292
+ {
1293
+ "epoch": 1.5625,
1294
+ "grad_norm": 0.2886108458042145,
1295
+ "learning_rate": 4.763620593732867e-06,
1296
+ "loss": 0.0521,
1297
+ "step": 184
1298
+ },
1299
+ {
1300
+ "epoch": 1.5710227272727273,
1301
+ "grad_norm": 0.23579014837741852,
1302
+ "learning_rate": 4.75805212923987e-06,
1303
+ "loss": 0.0551,
1304
+ "step": 185
1305
+ },
1306
+ {
1307
+ "epoch": 1.5795454545454546,
1308
+ "grad_norm": 0.2821877896785736,
1309
+ "learning_rate": 4.752422169756048e-06,
1310
+ "loss": 0.0494,
1311
+ "step": 186
1312
+ },
1313
+ {
1314
+ "epoch": 1.5880681818181817,
1315
+ "grad_norm": 0.28472304344177246,
1316
+ "learning_rate": 4.746730868605721e-06,
1317
+ "loss": 0.0524,
1318
+ "step": 187
1319
+ },
1320
+ {
1321
+ "epoch": 1.5965909090909092,
1322
+ "grad_norm": 0.30284813046455383,
1323
+ "learning_rate": 4.7409783807837654e-06,
1324
+ "loss": 0.0519,
1325
+ "step": 188
1326
+ },
1327
+ {
1328
+ "epoch": 1.6051136363636362,
1329
+ "grad_norm": 0.2568075656890869,
1330
+ "learning_rate": 4.735164862951395e-06,
1331
+ "loss": 0.0532,
1332
+ "step": 189
1333
+ },
1334
+ {
1335
+ "epoch": 1.6136363636363638,
1336
+ "grad_norm": 0.28199082612991333,
1337
+ "learning_rate": 4.729290473431892e-06,
1338
+ "loss": 0.0578,
1339
+ "step": 190
1340
+ },
1341
+ {
1342
+ "epoch": 1.6221590909090908,
1343
+ "grad_norm": 0.3818947970867157,
1344
+ "learning_rate": 4.723355372206297e-06,
1345
+ "loss": 0.0552,
1346
+ "step": 191
1347
+ },
1348
+ {
1349
+ "epoch": 1.6306818181818183,
1350
+ "grad_norm": 0.26314595341682434,
1351
+ "learning_rate": 4.717359720909053e-06,
1352
+ "loss": 0.0505,
1353
+ "step": 192
1354
+ },
1355
+ {
1356
+ "epoch": 1.6392045454545454,
1357
+ "grad_norm": 0.2616359293460846,
1358
+ "learning_rate": 4.7113036828235995e-06,
1359
+ "loss": 0.0511,
1360
+ "step": 193
1361
+ },
1362
+ {
1363
+ "epoch": 1.6477272727272727,
1364
+ "grad_norm": 0.33692309260368347,
1365
+ "learning_rate": 4.705187422877931e-06,
1366
+ "loss": 0.0554,
1367
+ "step": 194
1368
+ },
1369
+ {
1370
+ "epoch": 1.65625,
1371
+ "grad_norm": 0.30861467123031616,
1372
+ "learning_rate": 4.6990111076401e-06,
1373
+ "loss": 0.0485,
1374
+ "step": 195
1375
+ },
1376
+ {
1377
+ "epoch": 1.6647727272727273,
1378
+ "grad_norm": 0.251370370388031,
1379
+ "learning_rate": 4.692774905313687e-06,
1380
+ "loss": 0.0454,
1381
+ "step": 196
1382
+ },
1383
+ {
1384
+ "epoch": 1.6732954545454546,
1385
+ "grad_norm": 0.33005693554878235,
1386
+ "learning_rate": 4.686478985733212e-06,
1387
+ "loss": 0.0509,
1388
+ "step": 197
1389
+ },
1390
+ {
1391
+ "epoch": 1.6818181818181817,
1392
+ "grad_norm": 0.39172685146331787,
1393
+ "learning_rate": 4.68012352035952e-06,
1394
+ "loss": 0.0552,
1395
+ "step": 198
1396
+ },
1397
+ {
1398
+ "epoch": 1.6903409090909092,
1399
+ "grad_norm": 0.5113612413406372,
1400
+ "learning_rate": 4.673708682275097e-06,
1401
+ "loss": 0.0537,
1402
+ "step": 199
1403
+ },
1404
+ {
1405
+ "epoch": 1.6988636363636362,
1406
+ "grad_norm": 0.27660125494003296,
1407
+ "learning_rate": 4.667234646179368e-06,
1408
+ "loss": 0.0463,
1409
+ "step": 200
1410
+ },
1411
+ {
1412
+ "epoch": 1.7073863636363638,
1413
+ "grad_norm": 0.38760706782341003,
1414
+ "learning_rate": 4.660701588383939e-06,
1415
+ "loss": 0.0496,
1416
+ "step": 201
1417
+ },
1418
+ {
1419
+ "epoch": 1.7159090909090908,
1420
+ "grad_norm": 0.3487601578235626,
1421
+ "learning_rate": 4.654109686807787e-06,
1422
+ "loss": 0.0498,
1423
+ "step": 202
1424
+ },
1425
+ {
1426
+ "epoch": 1.7244318181818183,
1427
+ "grad_norm": 0.2964484393596649,
1428
+ "learning_rate": 4.647459120972422e-06,
1429
+ "loss": 0.0501,
1430
+ "step": 203
1431
+ },
1432
+ {
1433
+ "epoch": 1.7329545454545454,
1434
+ "grad_norm": 0.3318927586078644,
1435
+ "learning_rate": 4.640750071996995e-06,
1436
+ "loss": 0.0535,
1437
+ "step": 204
1438
+ },
1439
+ {
1440
+ "epoch": 1.7414772727272727,
1441
+ "grad_norm": 0.3152824640274048,
1442
+ "learning_rate": 4.633982722593367e-06,
1443
+ "loss": 0.0537,
1444
+ "step": 205
1445
+ },
1446
+ {
1447
+ "epoch": 1.75,
1448
+ "grad_norm": 0.3267940878868103,
1449
+ "learning_rate": 4.62715725706113e-06,
1450
+ "loss": 0.0434,
1451
+ "step": 206
1452
+ },
1453
+ {
1454
+ "epoch": 1.7585227272727273,
1455
+ "grad_norm": 0.32321202754974365,
1456
+ "learning_rate": 4.620273861282593e-06,
1457
+ "loss": 0.055,
1458
+ "step": 207
1459
+ },
1460
+ {
1461
+ "epoch": 1.7670454545454546,
1462
+ "grad_norm": 0.3501484990119934,
1463
+ "learning_rate": 4.613332722717714e-06,
1464
+ "loss": 0.0458,
1465
+ "step": 208
1466
+ },
1467
+ {
1468
+ "epoch": 1.7755681818181817,
1469
+ "grad_norm": 0.29107236862182617,
1470
+ "learning_rate": 4.606334030399e-06,
1471
+ "loss": 0.0496,
1472
+ "step": 209
1473
+ },
1474
+ {
1475
+ "epoch": 1.7840909090909092,
1476
+ "grad_norm": 0.3194922208786011,
1477
+ "learning_rate": 4.599277974926355e-06,
1478
+ "loss": 0.0532,
1479
+ "step": 210
1480
+ },
1481
+ {
1482
+ "epoch": 1.7926136363636362,
1483
+ "grad_norm": 0.2896871864795685,
1484
+ "learning_rate": 4.592164748461892e-06,
1485
+ "loss": 0.0482,
1486
+ "step": 211
1487
+ },
1488
+ {
1489
+ "epoch": 1.8011363636363638,
1490
+ "grad_norm": 0.2559587061405182,
1491
+ "learning_rate": 4.584994544724695e-06,
1492
+ "loss": 0.0474,
1493
+ "step": 212
1494
+ },
1495
+ {
1496
+ "epoch": 1.8096590909090908,
1497
+ "grad_norm": 0.3188915252685547,
1498
+ "learning_rate": 4.577767558985552e-06,
1499
+ "loss": 0.0454,
1500
+ "step": 213
1501
+ },
1502
+ {
1503
+ "epoch": 1.8181818181818183,
1504
+ "grad_norm": 0.31643223762512207,
1505
+ "learning_rate": 4.57048398806163e-06,
1506
+ "loss": 0.0503,
1507
+ "step": 214
1508
+ },
1509
+ {
1510
+ "epoch": 1.8267045454545454,
1511
+ "grad_norm": 0.3251838684082031,
1512
+ "learning_rate": 4.563144030311114e-06,
1513
+ "loss": 0.0484,
1514
+ "step": 215
1515
+ },
1516
+ {
1517
+ "epoch": 1.8352272727272727,
1518
+ "grad_norm": 0.26603975892066956,
1519
+ "learning_rate": 4.555747885627812e-06,
1520
+ "loss": 0.0442,
1521
+ "step": 216
1522
+ },
1523
+ {
1524
+ "epoch": 1.84375,
1525
+ "grad_norm": 0.4656536877155304,
1526
+ "learning_rate": 4.548295755435703e-06,
1527
+ "loss": 0.0428,
1528
+ "step": 217
1529
+ },
1530
+ {
1531
+ "epoch": 1.8522727272727273,
1532
+ "grad_norm": 0.32816216349601746,
1533
+ "learning_rate": 4.540787842683459e-06,
1534
+ "loss": 0.0455,
1535
+ "step": 218
1536
+ },
1537
+ {
1538
+ "epoch": 1.8607954545454546,
1539
+ "grad_norm": 0.31960371136665344,
1540
+ "learning_rate": 4.533224351838914e-06,
1541
+ "loss": 0.0454,
1542
+ "step": 219
1543
+ },
1544
+ {
1545
+ "epoch": 1.8693181818181817,
1546
+ "grad_norm": 0.33228686451911926,
1547
+ "learning_rate": 4.525605488883493e-06,
1548
+ "loss": 0.0413,
1549
+ "step": 220
1550
+ },
1551
+ {
1552
+ "epoch": 1.8778409090909092,
1553
+ "grad_norm": 0.271839439868927,
1554
+ "learning_rate": 4.517931461306609e-06,
1555
+ "loss": 0.0462,
1556
+ "step": 221
1557
+ },
1558
+ {
1559
+ "epoch": 1.8863636363636362,
1560
+ "grad_norm": 0.3333345949649811,
1561
+ "learning_rate": 4.510202478100008e-06,
1562
+ "loss": 0.0458,
1563
+ "step": 222
1564
+ },
1565
+ {
1566
+ "epoch": 1.8948863636363638,
1567
+ "grad_norm": 0.2923465371131897,
1568
+ "learning_rate": 4.502418749752076e-06,
1569
+ "loss": 0.0407,
1570
+ "step": 223
1571
+ },
1572
+ {
1573
+ "epoch": 1.9034090909090908,
1574
+ "grad_norm": 0.34088781476020813,
1575
+ "learning_rate": 4.494580488242109e-06,
1576
+ "loss": 0.0422,
1577
+ "step": 224
1578
+ },
1579
+ {
1580
+ "epoch": 1.9119318181818183,
1581
+ "grad_norm": 0.3535335958003998,
1582
+ "learning_rate": 4.486687907034544e-06,
1583
+ "loss": 0.0454,
1584
+ "step": 225
1585
+ },
1586
+ {
1587
+ "epoch": 1.9204545454545454,
1588
+ "grad_norm": 0.32973766326904297,
1589
+ "learning_rate": 4.478741221073136e-06,
1590
+ "loss": 0.0429,
1591
+ "step": 226
1592
+ },
1593
+ {
1594
+ "epoch": 1.9289772727272727,
1595
+ "grad_norm": 0.3509384095668793,
1596
+ "learning_rate": 4.470740646775113e-06,
1597
+ "loss": 0.0436,
1598
+ "step": 227
1599
+ },
1600
+ {
1601
+ "epoch": 1.9375,
1602
+ "grad_norm": 0.2931392788887024,
1603
+ "learning_rate": 4.462686402025277e-06,
1604
+ "loss": 0.0466,
1605
+ "step": 228
1606
+ },
1607
+ {
1608
+ "epoch": 1.9460227272727273,
1609
+ "grad_norm": 0.30415666103363037,
1610
+ "learning_rate": 4.454578706170075e-06,
1611
+ "loss": 0.0416,
1612
+ "step": 229
1613
+ },
1614
+ {
1615
+ "epoch": 1.9545454545454546,
1616
+ "grad_norm": 0.3433127701282501,
1617
+ "learning_rate": 4.446417780011618e-06,
1618
+ "loss": 0.0424,
1619
+ "step": 230
1620
+ },
1621
+ {
1622
+ "epoch": 1.9630681818181817,
1623
+ "grad_norm": 0.352159708738327,
1624
+ "learning_rate": 4.4382038458016764e-06,
1625
+ "loss": 0.0404,
1626
+ "step": 231
1627
+ },
1628
+ {
1629
+ "epoch": 1.9715909090909092,
1630
+ "grad_norm": 0.27307572960853577,
1631
+ "learning_rate": 4.42993712723562e-06,
1632
+ "loss": 0.0429,
1633
+ "step": 232
1634
+ },
1635
+ {
1636
+ "epoch": 1.9801136363636362,
1637
+ "grad_norm": 0.2662625312805176,
1638
+ "learning_rate": 4.4216178494463305e-06,
1639
+ "loss": 0.0425,
1640
+ "step": 233
1641
+ },
1642
+ {
1643
+ "epoch": 1.9886363636363638,
1644
+ "grad_norm": 0.29612037539482117,
1645
+ "learning_rate": 4.413246238998069e-06,
1646
+ "loss": 0.0463,
1647
+ "step": 234
1648
+ },
1649
+ {
1650
+ "epoch": 1.9971590909090908,
1651
+ "grad_norm": 0.33151501417160034,
1652
+ "learning_rate": 4.404822523880305e-06,
1653
+ "loss": 0.0441,
1654
+ "step": 235
1655
+ },
1656
+ {
1657
+ "epoch": 2.0,
1658
+ "grad_norm": 0.33151501417160034,
1659
+ "learning_rate": 4.396346933501508e-06,
1660
+ "loss": 0.0494,
1661
+ "step": 236
1662
+ },
1663
+ {
1664
+ "epoch": 2.008522727272727,
1665
+ "grad_norm": 0.8211016058921814,
1666
+ "learning_rate": 4.3878196986829015e-06,
1667
+ "loss": 0.0394,
1668
+ "step": 237
1669
+ },
1670
+ {
1671
+ "epoch": 2.0170454545454546,
1672
+ "grad_norm": 0.28689032793045044,
1673
+ "learning_rate": 4.379241051652174e-06,
1674
+ "loss": 0.0379,
1675
+ "step": 238
1676
+ },
1677
+ {
1678
+ "epoch": 2.0255681818181817,
1679
+ "grad_norm": 0.3541617691516876,
1680
+ "learning_rate": 4.370611226037155e-06,
1681
+ "loss": 0.0363,
1682
+ "step": 239
1683
+ },
1684
+ {
1685
+ "epoch": 2.034090909090909,
1686
+ "grad_norm": 0.32633963227272034,
1687
+ "learning_rate": 4.361930456859455e-06,
1688
+ "loss": 0.0403,
1689
+ "step": 240
1690
+ },
1691
+ {
1692
+ "epoch": 2.0426136363636362,
1693
+ "grad_norm": 0.30213046073913574,
1694
+ "learning_rate": 4.353198980528063e-06,
1695
+ "loss": 0.0393,
1696
+ "step": 241
1697
+ },
1698
+ {
1699
+ "epoch": 2.0511363636363638,
1700
+ "grad_norm": 0.3540145456790924,
1701
+ "learning_rate": 4.3444170348329095e-06,
1702
+ "loss": 0.035,
1703
+ "step": 242
1704
+ },
1705
+ {
1706
+ "epoch": 2.059659090909091,
1707
+ "grad_norm": 0.2990974485874176,
1708
+ "learning_rate": 4.335584858938388e-06,
1709
+ "loss": 0.0346,
1710
+ "step": 243
1711
+ },
1712
+ {
1713
+ "epoch": 2.0681818181818183,
1714
+ "grad_norm": 0.4325309991836548,
1715
+ "learning_rate": 4.326702693376844e-06,
1716
+ "loss": 0.0351,
1717
+ "step": 244
1718
+ },
1719
+ {
1720
+ "epoch": 2.0767045454545454,
1721
+ "grad_norm": 0.31114035844802856,
1722
+ "learning_rate": 4.317770780042025e-06,
1723
+ "loss": 0.0341,
1724
+ "step": 245
1725
+ },
1726
+ {
1727
+ "epoch": 2.085227272727273,
1728
+ "grad_norm": 0.30371102690696716,
1729
+ "learning_rate": 4.308789362182492e-06,
1730
+ "loss": 0.0373,
1731
+ "step": 246
1732
+ },
1733
+ {
1734
+ "epoch": 2.09375,
1735
+ "grad_norm": 0.40639927983283997,
1736
+ "learning_rate": 4.2997586843949905e-06,
1737
+ "loss": 0.0333,
1738
+ "step": 247
1739
+ },
1740
+ {
1741
+ "epoch": 2.102272727272727,
1742
+ "grad_norm": 0.33129221200942993,
1743
+ "learning_rate": 4.290678992617797e-06,
1744
+ "loss": 0.0323,
1745
+ "step": 248
1746
+ },
1747
+ {
1748
+ "epoch": 2.1107954545454546,
1749
+ "grad_norm": 0.3084648549556732,
1750
+ "learning_rate": 4.28155053412402e-06,
1751
+ "loss": 0.0355,
1752
+ "step": 249
1753
+ },
1754
+ {
1755
+ "epoch": 2.1193181818181817,
1756
+ "grad_norm": 0.3743211328983307,
1757
+ "learning_rate": 4.2723735575148585e-06,
1758
+ "loss": 0.0367,
1759
+ "step": 250
1760
+ },
1761
+ {
1762
+ "epoch": 2.127840909090909,
1763
+ "grad_norm": 0.354165256023407,
1764
+ "learning_rate": 4.263148312712838e-06,
1765
+ "loss": 0.0293,
1766
+ "step": 251
1767
+ },
1768
+ {
1769
+ "epoch": 2.1363636363636362,
1770
+ "grad_norm": 0.37607696652412415,
1771
+ "learning_rate": 4.253875050955005e-06,
1772
+ "loss": 0.0318,
1773
+ "step": 252
1774
+ },
1775
+ {
1776
+ "epoch": 2.1448863636363638,
1777
+ "grad_norm": 0.33469027280807495,
1778
+ "learning_rate": 4.2445540247860805e-06,
1779
+ "loss": 0.038,
1780
+ "step": 253
1781
+ },
1782
+ {
1783
+ "epoch": 2.153409090909091,
1784
+ "grad_norm": 0.32785046100616455,
1785
+ "learning_rate": 4.2351854880515856e-06,
1786
+ "loss": 0.0303,
1787
+ "step": 254
1788
+ },
1789
+ {
1790
+ "epoch": 2.1619318181818183,
1791
+ "grad_norm": 0.41809284687042236,
1792
+ "learning_rate": 4.2257696958909255e-06,
1793
+ "loss": 0.0339,
1794
+ "step": 255
1795
+ },
1796
+ {
1797
+ "epoch": 2.1704545454545454,
1798
+ "grad_norm": 0.3557347059249878,
1799
+ "learning_rate": 4.216306904730448e-06,
1800
+ "loss": 0.0347,
1801
+ "step": 256
1802
+ },
1803
+ {
1804
+ "epoch": 2.178977272727273,
1805
+ "grad_norm": 0.3383937478065491,
1806
+ "learning_rate": 4.206797372276447e-06,
1807
+ "loss": 0.0384,
1808
+ "step": 257
1809
+ },
1810
+ {
1811
+ "epoch": 2.1875,
1812
+ "grad_norm": 0.339476615190506,
1813
+ "learning_rate": 4.197241357508159e-06,
1814
+ "loss": 0.0351,
1815
+ "step": 258
1816
+ },
1817
+ {
1818
+ "epoch": 2.196022727272727,
1819
+ "grad_norm": 0.4095429480075836,
1820
+ "learning_rate": 4.187639120670702e-06,
1821
+ "loss": 0.0339,
1822
+ "step": 259
1823
+ },
1824
+ {
1825
+ "epoch": 2.2045454545454546,
1826
+ "grad_norm": 0.3317232131958008,
1827
+ "learning_rate": 4.177990923267986e-06,
1828
+ "loss": 0.0363,
1829
+ "step": 260
1830
+ },
1831
+ {
1832
+ "epoch": 2.2130681818181817,
1833
+ "grad_norm": 0.4111759066581726,
1834
+ "learning_rate": 4.168297028055599e-06,
1835
+ "loss": 0.0369,
1836
+ "step": 261
1837
+ },
1838
+ {
1839
+ "epoch": 2.221590909090909,
1840
+ "grad_norm": 0.4489494860172272,
1841
+ "learning_rate": 4.158557699033644e-06,
1842
+ "loss": 0.0312,
1843
+ "step": 262
1844
+ },
1845
+ {
1846
+ "epoch": 2.2301136363636362,
1847
+ "grad_norm": 0.3993566930294037,
1848
+ "learning_rate": 4.148773201439553e-06,
1849
+ "loss": 0.0306,
1850
+ "step": 263
1851
+ },
1852
+ {
1853
+ "epoch": 2.2386363636363638,
1854
+ "grad_norm": 0.32782599329948425,
1855
+ "learning_rate": 4.138943801740865e-06,
1856
+ "loss": 0.0329,
1857
+ "step": 264
1858
+ },
1859
+ {
1860
+ "epoch": 2.247159090909091,
1861
+ "grad_norm": 0.26705044507980347,
1862
+ "learning_rate": 4.129069767627963e-06,
1863
+ "loss": 0.0341,
1864
+ "step": 265
1865
+ },
1866
+ {
1867
+ "epoch": 2.2556818181818183,
1868
+ "grad_norm": 0.32875779271125793,
1869
+ "learning_rate": 4.119151368006793e-06,
1870
+ "loss": 0.0301,
1871
+ "step": 266
1872
+ },
1873
+ {
1874
+ "epoch": 2.2642045454545454,
1875
+ "grad_norm": 0.42058393359184265,
1876
+ "learning_rate": 4.109188872991529e-06,
1877
+ "loss": 0.0346,
1878
+ "step": 267
1879
+ },
1880
+ {
1881
+ "epoch": 2.2727272727272725,
1882
+ "grad_norm": 0.44719040393829346,
1883
+ "learning_rate": 4.099182553897228e-06,
1884
+ "loss": 0.0297,
1885
+ "step": 268
1886
+ },
1887
+ {
1888
+ "epoch": 2.28125,
1889
+ "grad_norm": 0.35350751876831055,
1890
+ "learning_rate": 4.089132683232437e-06,
1891
+ "loss": 0.0285,
1892
+ "step": 269
1893
+ },
1894
+ {
1895
+ "epoch": 2.2897727272727275,
1896
+ "grad_norm": 0.5518717169761658,
1897
+ "learning_rate": 4.0790395346917674e-06,
1898
+ "loss": 0.0333,
1899
+ "step": 270
1900
+ },
1901
+ {
1902
+ "epoch": 2.2982954545454546,
1903
+ "grad_norm": 0.4138563573360443,
1904
+ "learning_rate": 4.068903383148448e-06,
1905
+ "loss": 0.0385,
1906
+ "step": 271
1907
+ },
1908
+ {
1909
+ "epoch": 2.3068181818181817,
1910
+ "grad_norm": 0.406649112701416,
1911
+ "learning_rate": 4.058724504646834e-06,
1912
+ "loss": 0.0298,
1913
+ "step": 272
1914
+ },
1915
+ {
1916
+ "epoch": 2.315340909090909,
1917
+ "grad_norm": 0.3303579092025757,
1918
+ "learning_rate": 4.0485031763948935e-06,
1919
+ "loss": 0.0339,
1920
+ "step": 273
1921
+ },
1922
+ {
1923
+ "epoch": 2.3238636363636362,
1924
+ "grad_norm": 0.35349851846694946,
1925
+ "learning_rate": 4.038239676756654e-06,
1926
+ "loss": 0.0308,
1927
+ "step": 274
1928
+ },
1929
+ {
1930
+ "epoch": 2.3323863636363638,
1931
+ "grad_norm": 0.4232323467731476,
1932
+ "learning_rate": 4.027934285244624e-06,
1933
+ "loss": 0.0307,
1934
+ "step": 275
1935
+ },
1936
+ {
1937
+ "epoch": 2.340909090909091,
1938
+ "grad_norm": 0.35764482617378235,
1939
+ "learning_rate": 4.017587282512181e-06,
1940
+ "loss": 0.0296,
1941
+ "step": 276
1942
+ },
1943
+ {
1944
+ "epoch": 2.3494318181818183,
1945
+ "grad_norm": 0.3608599603176117,
1946
+ "learning_rate": 4.007198950345929e-06,
1947
+ "loss": 0.0346,
1948
+ "step": 277
1949
+ },
1950
+ {
1951
+ "epoch": 2.3579545454545454,
1952
+ "grad_norm": 0.530514121055603,
1953
+ "learning_rate": 3.996769571658022e-06,
1954
+ "loss": 0.0293,
1955
+ "step": 278
1956
+ },
1957
+ {
1958
+ "epoch": 2.3664772727272725,
1959
+ "grad_norm": 0.3813832998275757,
1960
+ "learning_rate": 3.9862994304784585e-06,
1961
+ "loss": 0.0279,
1962
+ "step": 279
1963
+ },
1964
+ {
1965
+ "epoch": 2.375,
1966
+ "grad_norm": 0.48923951387405396,
1967
+ "learning_rate": 3.975788811947351e-06,
1968
+ "loss": 0.0284,
1969
+ "step": 280
1970
+ },
1971
+ {
1972
+ "epoch": 2.3835227272727275,
1973
+ "grad_norm": 0.3939470946788788,
1974
+ "learning_rate": 3.965238002307156e-06,
1975
+ "loss": 0.0315,
1976
+ "step": 281
1977
+ },
1978
+ {
1979
+ "epoch": 2.3920454545454546,
1980
+ "grad_norm": 0.2854841649532318,
1981
+ "learning_rate": 3.9546472888948825e-06,
1982
+ "loss": 0.0288,
1983
+ "step": 282
1984
+ },
1985
+ {
1986
+ "epoch": 2.4005681818181817,
1987
+ "grad_norm": 0.39930015802383423,
1988
+ "learning_rate": 3.944016960134262e-06,
1989
+ "loss": 0.0304,
1990
+ "step": 283
1991
+ },
1992
+ {
1993
+ "epoch": 2.409090909090909,
1994
+ "grad_norm": 0.4333280026912689,
1995
+ "learning_rate": 3.933347305527898e-06,
1996
+ "loss": 0.0282,
1997
+ "step": 284
1998
+ },
1999
+ {
2000
+ "epoch": 2.4176136363636362,
2001
+ "grad_norm": 0.34740084409713745,
2002
+ "learning_rate": 3.922638615649381e-06,
2003
+ "loss": 0.0313,
2004
+ "step": 285
2005
+ },
2006
+ {
2007
+ "epoch": 2.4261363636363638,
2008
+ "grad_norm": 0.4125167727470398,
2009
+ "learning_rate": 3.911891182135371e-06,
2010
+ "loss": 0.0301,
2011
+ "step": 286
2012
+ },
2013
+ {
2014
+ "epoch": 2.434659090909091,
2015
+ "grad_norm": 0.350911945104599,
2016
+ "learning_rate": 3.9011052976776635e-06,
2017
+ "loss": 0.0255,
2018
+ "step": 287
2019
+ },
2020
+ {
2021
+ "epoch": 2.4431818181818183,
2022
+ "grad_norm": 0.2700820863246918,
2023
+ "learning_rate": 3.890281256015207e-06,
2024
+ "loss": 0.0296,
2025
+ "step": 288
2026
+ },
2027
+ {
2028
+ "epoch": 2.4517045454545454,
2029
+ "grad_norm": 0.3823346197605133,
2030
+ "learning_rate": 3.879419351926115e-06,
2031
+ "loss": 0.0279,
2032
+ "step": 289
2033
+ },
2034
+ {
2035
+ "epoch": 2.4602272727272725,
2036
+ "grad_norm": 0.4576987326145172,
2037
+ "learning_rate": 3.868519881219631e-06,
2038
+ "loss": 0.0298,
2039
+ "step": 290
2040
+ },
2041
+ {
2042
+ "epoch": 2.46875,
2043
+ "grad_norm": 0.319879412651062,
2044
+ "learning_rate": 3.857583140728075e-06,
2045
+ "loss": 0.0313,
2046
+ "step": 291
2047
+ },
2048
+ {
2049
+ "epoch": 2.4772727272727275,
2050
+ "grad_norm": 0.36349669098854065,
2051
+ "learning_rate": 3.8466094282987575e-06,
2052
+ "loss": 0.0266,
2053
+ "step": 292
2054
+ },
2055
+ {
2056
+ "epoch": 2.4857954545454546,
2057
+ "grad_norm": 0.3291328251361847,
2058
+ "learning_rate": 3.835599042785872e-06,
2059
+ "loss": 0.0292,
2060
+ "step": 293
2061
+ },
2062
+ {
2063
+ "epoch": 2.4943181818181817,
2064
+ "grad_norm": 0.36020705103874207,
2065
+ "learning_rate": 3.824552284042351e-06,
2066
+ "loss": 0.0295,
2067
+ "step": 294
2068
+ },
2069
+ {
2070
+ "epoch": 2.502840909090909,
2071
+ "grad_norm": 0.3137778341770172,
2072
+ "learning_rate": 3.8134694529117045e-06,
2073
+ "loss": 0.0244,
2074
+ "step": 295
2075
+ },
2076
+ {
2077
+ "epoch": 2.5113636363636362,
2078
+ "grad_norm": 0.36281993985176086,
2079
+ "learning_rate": 3.802350851219826e-06,
2080
+ "loss": 0.0272,
2081
+ "step": 296
2082
+ },
2083
+ {
2084
+ "epoch": 2.5198863636363638,
2085
+ "grad_norm": 0.3342132270336151,
2086
+ "learning_rate": 3.79119678176677e-06,
2087
+ "loss": 0.0288,
2088
+ "step": 297
2089
+ },
2090
+ {
2091
+ "epoch": 2.528409090909091,
2092
+ "grad_norm": 0.45444363355636597,
2093
+ "learning_rate": 3.7800075483185073e-06,
2094
+ "loss": 0.0282,
2095
+ "step": 298
2096
+ },
2097
+ {
2098
+ "epoch": 2.5369318181818183,
2099
+ "grad_norm": 0.4871613085269928,
2100
+ "learning_rate": 3.7687834555986537e-06,
2101
+ "loss": 0.0334,
2102
+ "step": 299
2103
+ },
2104
+ {
2105
+ "epoch": 2.5454545454545454,
2106
+ "grad_norm": 0.49083876609802246,
2107
+ "learning_rate": 3.7575248092801686e-06,
2108
+ "loss": 0.0229,
2109
+ "step": 300
2110
+ },
2111
+ {
2112
+ "epoch": 2.5539772727272725,
2113
+ "grad_norm": 0.3823300004005432,
2114
+ "learning_rate": 3.7462319159770344e-06,
2115
+ "loss": 0.023,
2116
+ "step": 301
2117
+ },
2118
+ {
2119
+ "epoch": 2.5625,
2120
+ "grad_norm": 0.38027557730674744,
2121
+ "learning_rate": 3.734905083235901e-06,
2122
+ "loss": 0.025,
2123
+ "step": 302
2124
+ },
2125
+ {
2126
+ "epoch": 2.5710227272727275,
2127
+ "grad_norm": 0.42017418146133423,
2128
+ "learning_rate": 3.723544619527714e-06,
2129
+ "loss": 0.03,
2130
+ "step": 303
2131
+ },
2132
+ {
2133
+ "epoch": 2.5795454545454546,
2134
+ "grad_norm": 0.4550953209400177,
2135
+ "learning_rate": 3.712150834239313e-06,
2136
+ "loss": 0.0301,
2137
+ "step": 304
2138
+ },
2139
+ {
2140
+ "epoch": 2.5880681818181817,
2141
+ "grad_norm": 0.4527112543582916,
2142
+ "learning_rate": 3.7007240376650054e-06,
2143
+ "loss": 0.0235,
2144
+ "step": 305
2145
+ },
2146
+ {
2147
+ "epoch": 2.596590909090909,
2148
+ "grad_norm": 0.36789700388908386,
2149
+ "learning_rate": 3.6892645409981166e-06,
2150
+ "loss": 0.0231,
2151
+ "step": 306
2152
+ },
2153
+ {
2154
+ "epoch": 2.6051136363636362,
2155
+ "grad_norm": 0.6067688465118408,
2156
+ "learning_rate": 3.6777726563225147e-06,
2157
+ "loss": 0.0256,
2158
+ "step": 307
2159
+ },
2160
+ {
2161
+ "epoch": 2.6136363636363638,
2162
+ "grad_norm": 0.35337361693382263,
2163
+ "learning_rate": 3.6662486966041104e-06,
2164
+ "loss": 0.0244,
2165
+ "step": 308
2166
+ },
2167
+ {
2168
+ "epoch": 2.622159090909091,
2169
+ "grad_norm": 0.4097234010696411,
2170
+ "learning_rate": 3.654692975682336e-06,
2171
+ "loss": 0.0262,
2172
+ "step": 309
2173
+ },
2174
+ {
2175
+ "epoch": 2.6306818181818183,
2176
+ "grad_norm": 0.5017573833465576,
2177
+ "learning_rate": 3.6431058082615966e-06,
2178
+ "loss": 0.0245,
2179
+ "step": 310
2180
+ },
2181
+ {
2182
+ "epoch": 2.6392045454545454,
2183
+ "grad_norm": 0.40696877241134644,
2184
+ "learning_rate": 3.6314875099027e-06,
2185
+ "loss": 0.023,
2186
+ "step": 311
2187
+ },
2188
+ {
2189
+ "epoch": 2.6477272727272725,
2190
+ "grad_norm": 0.35261037945747375,
2191
+ "learning_rate": 3.619838397014263e-06,
2192
+ "loss": 0.0232,
2193
+ "step": 312
2194
+ },
2195
+ {
2196
+ "epoch": 2.65625,
2197
+ "grad_norm": 0.3590414524078369,
2198
+ "learning_rate": 3.6081587868440944e-06,
2199
+ "loss": 0.0242,
2200
+ "step": 313
2201
+ },
2202
+ {
2203
+ "epoch": 2.6647727272727275,
2204
+ "grad_norm": 0.37196213006973267,
2205
+ "learning_rate": 3.5964489974705553e-06,
2206
+ "loss": 0.0222,
2207
+ "step": 314
2208
+ },
2209
+ {
2210
+ "epoch": 2.6732954545454546,
2211
+ "grad_norm": 0.3922335207462311,
2212
+ "learning_rate": 3.5847093477938955e-06,
2213
+ "loss": 0.0236,
2214
+ "step": 315
2215
+ },
2216
+ {
2217
+ "epoch": 2.6818181818181817,
2218
+ "grad_norm": 0.3914213180541992,
2219
+ "learning_rate": 3.5729401575275724e-06,
2220
+ "loss": 0.025,
2221
+ "step": 316
2222
+ },
2223
+ {
2224
+ "epoch": 2.690340909090909,
2225
+ "grad_norm": 0.4622058570384979,
2226
+ "learning_rate": 3.561141747189538e-06,
2227
+ "loss": 0.0252,
2228
+ "step": 317
2229
+ },
2230
+ {
2231
+ "epoch": 2.6988636363636362,
2232
+ "grad_norm": 0.5096047520637512,
2233
+ "learning_rate": 3.5493144380935155e-06,
2234
+ "loss": 0.0242,
2235
+ "step": 318
2236
+ },
2237
+ {
2238
+ "epoch": 2.7073863636363638,
2239
+ "grad_norm": 0.574635922908783,
2240
+ "learning_rate": 3.537458552340247e-06,
2241
+ "loss": 0.0224,
2242
+ "step": 319
2243
+ },
2244
+ {
2245
+ "epoch": 2.715909090909091,
2246
+ "grad_norm": 0.6871437430381775,
2247
+ "learning_rate": 3.5255744128087175e-06,
2248
+ "loss": 0.0271,
2249
+ "step": 320
2250
+ },
2251
+ {
2252
+ "epoch": 2.7244318181818183,
2253
+ "grad_norm": 0.48472297191619873,
2254
+ "learning_rate": 3.5136623431473704e-06,
2255
+ "loss": 0.0246,
2256
+ "step": 321
2257
+ },
2258
+ {
2259
+ "epoch": 2.7329545454545454,
2260
+ "grad_norm": 0.5469712018966675,
2261
+ "learning_rate": 3.501722667765286e-06,
2262
+ "loss": 0.0252,
2263
+ "step": 322
2264
+ },
2265
+ {
2266
+ "epoch": 2.7414772727272725,
2267
+ "grad_norm": 0.4378078877925873,
2268
+ "learning_rate": 3.489755711823348e-06,
2269
+ "loss": 0.0277,
2270
+ "step": 323
2271
+ },
2272
+ {
2273
+ "epoch": 2.75,
2274
+ "grad_norm": 0.4401378035545349,
2275
+ "learning_rate": 3.47776180122539e-06,
2276
+ "loss": 0.0235,
2277
+ "step": 324
2278
+ },
2279
+ {
2280
+ "epoch": 2.7585227272727275,
2281
+ "grad_norm": 0.6726410388946533,
2282
+ "learning_rate": 3.4657412626093204e-06,
2283
+ "loss": 0.0207,
2284
+ "step": 325
2285
+ },
2286
+ {
2287
+ "epoch": 2.7670454545454546,
2288
+ "grad_norm": 0.3676104247570038,
2289
+ "learning_rate": 3.4536944233382248e-06,
2290
+ "loss": 0.0196,
2291
+ "step": 326
2292
+ },
2293
+ {
2294
+ "epoch": 2.7755681818181817,
2295
+ "grad_norm": 0.4041842222213745,
2296
+ "learning_rate": 3.4416216114914493e-06,
2297
+ "loss": 0.0275,
2298
+ "step": 327
2299
+ },
2300
+ {
2301
+ "epoch": 2.784090909090909,
2302
+ "grad_norm": 0.7207781076431274,
2303
+ "learning_rate": 3.429523155855672e-06,
2304
+ "loss": 0.0243,
2305
+ "step": 328
2306
+ },
2307
+ {
2308
+ "epoch": 2.7926136363636362,
2309
+ "grad_norm": 0.6210029125213623,
2310
+ "learning_rate": 3.417399385915941e-06,
2311
+ "loss": 0.0207,
2312
+ "step": 329
2313
+ },
2314
+ {
2315
+ "epoch": 2.8011363636363638,
2316
+ "grad_norm": 0.49545082449913025,
2317
+ "learning_rate": 3.405250631846708e-06,
2318
+ "loss": 0.0243,
2319
+ "step": 330
2320
+ },
2321
+ {
2322
+ "epoch": 2.809659090909091,
2323
+ "grad_norm": 0.5395477414131165,
2324
+ "learning_rate": 3.393077224502832e-06,
2325
+ "loss": 0.0227,
2326
+ "step": 331
2327
+ },
2328
+ {
2329
+ "epoch": 2.8181818181818183,
2330
+ "grad_norm": 0.4645976126194,
2331
+ "learning_rate": 3.3808794954105716e-06,
2332
+ "loss": 0.0192,
2333
+ "step": 332
2334
+ },
2335
+ {
2336
+ "epoch": 2.8267045454545454,
2337
+ "grad_norm": 0.3877430558204651,
2338
+ "learning_rate": 3.368657776758555e-06,
2339
+ "loss": 0.0219,
2340
+ "step": 333
2341
+ },
2342
+ {
2343
+ "epoch": 2.8352272727272725,
2344
+ "grad_norm": 0.5611310005187988,
2345
+ "learning_rate": 3.3564124013887324e-06,
2346
+ "loss": 0.0241,
2347
+ "step": 334
2348
+ },
2349
+ {
2350
+ "epoch": 2.84375,
2351
+ "grad_norm": 0.4889032244682312,
2352
+ "learning_rate": 3.3441437027873135e-06,
2353
+ "loss": 0.0181,
2354
+ "step": 335
2355
+ },
2356
+ {
2357
+ "epoch": 2.8522727272727275,
2358
+ "grad_norm": 0.47838833928108215,
2359
+ "learning_rate": 3.331852015075685e-06,
2360
+ "loss": 0.0192,
2361
+ "step": 336
2362
+ },
2363
+ {
2364
+ "epoch": 2.8607954545454546,
2365
+ "grad_norm": 0.4922442138195038,
2366
+ "learning_rate": 3.3195376730013107e-06,
2367
+ "loss": 0.0197,
2368
+ "step": 337
2369
+ },
2370
+ {
2371
+ "epoch": 2.8693181818181817,
2372
+ "grad_norm": 0.3534923493862152,
2373
+ "learning_rate": 3.3072010119286156e-06,
2374
+ "loss": 0.0221,
2375
+ "step": 338
2376
+ },
2377
+ {
2378
+ "epoch": 2.877840909090909,
2379
+ "grad_norm": 0.4476766884326935,
2380
+ "learning_rate": 3.294842367829851e-06,
2381
+ "loss": 0.0167,
2382
+ "step": 339
2383
+ },
2384
+ {
2385
+ "epoch": 2.8863636363636362,
2386
+ "grad_norm": 0.4423317313194275,
2387
+ "learning_rate": 3.2824620772759475e-06,
2388
+ "loss": 0.0209,
2389
+ "step": 340
2390
+ },
2391
+ {
2392
+ "epoch": 2.8948863636363638,
2393
+ "grad_norm": 0.3954547345638275,
2394
+ "learning_rate": 3.270060477427347e-06,
2395
+ "loss": 0.0209,
2396
+ "step": 341
2397
+ },
2398
+ {
2399
+ "epoch": 2.903409090909091,
2400
+ "grad_norm": 0.47028860449790955,
2401
+ "learning_rate": 3.257637906024822e-06,
2402
+ "loss": 0.0189,
2403
+ "step": 342
2404
+ },
2405
+ {
2406
+ "epoch": 2.9119318181818183,
2407
+ "grad_norm": 0.42390963435173035,
2408
+ "learning_rate": 3.2451947013802746e-06,
2409
+ "loss": 0.015,
2410
+ "step": 343
2411
+ },
2412
+ {
2413
+ "epoch": 2.9204545454545454,
2414
+ "grad_norm": 0.4463144838809967,
2415
+ "learning_rate": 3.2327312023675287e-06,
2416
+ "loss": 0.0177,
2417
+ "step": 344
2418
+ },
2419
+ {
2420
+ "epoch": 2.9289772727272725,
2421
+ "grad_norm": 0.43879300355911255,
2422
+ "learning_rate": 3.2202477484130947e-06,
2423
+ "loss": 0.0145,
2424
+ "step": 345
2425
+ },
2426
+ {
2427
+ "epoch": 2.9375,
2428
+ "grad_norm": 0.36773353815078735,
2429
+ "learning_rate": 3.20774467948693e-06,
2430
+ "loss": 0.0186,
2431
+ "step": 346
2432
+ },
2433
+ {
2434
+ "epoch": 2.9460227272727275,
2435
+ "grad_norm": 0.45109617710113525,
2436
+ "learning_rate": 3.19522233609318e-06,
2437
+ "loss": 0.0153,
2438
+ "step": 347
2439
+ },
2440
+ {
2441
+ "epoch": 2.9545454545454546,
2442
+ "grad_norm": 0.48943161964416504,
2443
+ "learning_rate": 3.1826810592609036e-06,
2444
+ "loss": 0.0186,
2445
+ "step": 348
2446
+ },
2447
+ {
2448
+ "epoch": 2.9630681818181817,
2449
+ "grad_norm": 0.5407648682594299,
2450
+ "learning_rate": 3.170121190534785e-06,
2451
+ "loss": 0.0173,
2452
+ "step": 349
2453
+ },
2454
+ {
2455
+ "epoch": 2.971590909090909,
2456
+ "grad_norm": 0.5848131775856018,
2457
+ "learning_rate": 3.157543071965835e-06,
2458
+ "loss": 0.013,
2459
+ "step": 350
2460
+ },
2461
+ {
2462
+ "epoch": 2.9801136363636362,
2463
+ "grad_norm": 0.46287980675697327,
2464
+ "learning_rate": 3.1449470461020738e-06,
2465
+ "loss": 0.0194,
2466
+ "step": 351
2467
+ },
2468
+ {
2469
+ "epoch": 2.9886363636363638,
2470
+ "grad_norm": 0.655013382434845,
2471
+ "learning_rate": 3.132333455979202e-06,
2472
+ "loss": 0.0176,
2473
+ "step": 352
2474
+ },
2475
+ {
2476
+ "epoch": 2.997159090909091,
2477
+ "grad_norm": 0.7326741218566895,
2478
+ "learning_rate": 3.11970264511126e-06,
2479
+ "loss": 0.0154,
2480
+ "step": 353
2481
+ },
2482
+ {
2483
+ "epoch": 3.0,
2484
+ "grad_norm": 0.9463636875152588,
2485
+ "learning_rate": 3.107054957481271e-06,
2486
+ "loss": 0.0142,
2487
+ "step": 354
2488
+ },
2489
+ {
2490
+ "epoch": 3.008522727272727,
2491
+ "grad_norm": 0.8092179894447327,
2492
+ "learning_rate": 3.0943907375318733e-06,
2493
+ "loss": 0.0178,
2494
+ "step": 355
2495
+ },
2496
+ {
2497
+ "epoch": 3.0170454545454546,
2498
+ "grad_norm": 0.43350231647491455,
2499
+ "learning_rate": 3.0817103301559422e-06,
2500
+ "loss": 0.0163,
2501
+ "step": 356
2502
+ },
2503
+ {
2504
+ "epoch": 3.0255681818181817,
2505
+ "grad_norm": 0.3987981379032135,
2506
+ "learning_rate": 3.069014080687195e-06,
2507
+ "loss": 0.0153,
2508
+ "step": 357
2509
+ },
2510
+ {
2511
+ "epoch": 3.034090909090909,
2512
+ "grad_norm": 0.3733361065387726,
2513
+ "learning_rate": 3.056302334890786e-06,
2514
+ "loss": 0.012,
2515
+ "step": 358
2516
+ },
2517
+ {
2518
+ "epoch": 3.0426136363636362,
2519
+ "grad_norm": 0.5638909339904785,
2520
+ "learning_rate": 3.043575438953893e-06,
2521
+ "loss": 0.0128,
2522
+ "step": 359
2523
+ },
2524
+ {
2525
+ "epoch": 3.0511363636363638,
2526
+ "grad_norm": 0.5339016914367676,
2527
+ "learning_rate": 3.030833739476285e-06,
2528
+ "loss": 0.0142,
2529
+ "step": 360
2530
+ },
2531
+ {
2532
+ "epoch": 3.059659090909091,
2533
+ "grad_norm": 0.5681723356246948,
2534
+ "learning_rate": 3.018077583460886e-06,
2535
+ "loss": 0.0144,
2536
+ "step": 361
2537
+ },
2538
+ {
2539
+ "epoch": 3.0681818181818183,
2540
+ "grad_norm": 0.5017768144607544,
2541
+ "learning_rate": 3.0053073183043257e-06,
2542
+ "loss": 0.0137,
2543
+ "step": 362
2544
+ },
2545
+ {
2546
+ "epoch": 3.0767045454545454,
2547
+ "grad_norm": 1.0726104974746704,
2548
+ "learning_rate": 2.9925232917874764e-06,
2549
+ "loss": 0.013,
2550
+ "step": 363
2551
+ },
2552
+ {
2553
+ "epoch": 3.085227272727273,
2554
+ "grad_norm": 0.5261201858520508,
2555
+ "learning_rate": 2.979725852065981e-06,
2556
+ "loss": 0.0122,
2557
+ "step": 364
2558
+ },
2559
+ {
2560
+ "epoch": 3.09375,
2561
+ "grad_norm": 0.5007719993591309,
2562
+ "learning_rate": 2.966915347660775e-06,
2563
+ "loss": 0.0123,
2564
+ "step": 365
2565
+ },
2566
+ {
2567
+ "epoch": 3.102272727272727,
2568
+ "grad_norm": 0.5095813870429993,
2569
+ "learning_rate": 2.9540921274485913e-06,
2570
+ "loss": 0.0133,
2571
+ "step": 366
2572
+ },
2573
+ {
2574
+ "epoch": 3.1107954545454546,
2575
+ "grad_norm": 0.4466419816017151,
2576
+ "learning_rate": 2.941256540652462e-06,
2577
+ "loss": 0.0094,
2578
+ "step": 367
2579
+ },
2580
+ {
2581
+ "epoch": 3.1193181818181817,
2582
+ "grad_norm": 0.9082984328269958,
2583
+ "learning_rate": 2.9284089368322044e-06,
2584
+ "loss": 0.0155,
2585
+ "step": 368
2586
+ },
2587
+ {
2588
+ "epoch": 3.127840909090909,
2589
+ "grad_norm": 0.4232461750507355,
2590
+ "learning_rate": 2.915549665874905e-06,
2591
+ "loss": 0.0103,
2592
+ "step": 369
2593
+ },
2594
+ {
2595
+ "epoch": 3.1363636363636362,
2596
+ "grad_norm": 0.736997663974762,
2597
+ "learning_rate": 2.9026790779853877e-06,
2598
+ "loss": 0.0153,
2599
+ "step": 370
2600
+ },
2601
+ {
2602
+ "epoch": 3.1448863636363638,
2603
+ "grad_norm": 0.5333961844444275,
2604
+ "learning_rate": 2.8897975236766784e-06,
2605
+ "loss": 0.0122,
2606
+ "step": 371
2607
+ },
2608
+ {
2609
+ "epoch": 3.153409090909091,
2610
+ "grad_norm": 0.610040545463562,
2611
+ "learning_rate": 2.876905353760459e-06,
2612
+ "loss": 0.0121,
2613
+ "step": 372
2614
+ },
2615
+ {
2616
+ "epoch": 3.1619318181818183,
2617
+ "grad_norm": 0.3609560430049896,
2618
+ "learning_rate": 2.864002919337513e-06,
2619
+ "loss": 0.0106,
2620
+ "step": 373
2621
+ },
2622
+ {
2623
+ "epoch": 3.1704545454545454,
2624
+ "grad_norm": 0.5127677917480469,
2625
+ "learning_rate": 2.8510905717881615e-06,
2626
+ "loss": 0.0146,
2627
+ "step": 374
2628
+ },
2629
+ {
2630
+ "epoch": 3.178977272727273,
2631
+ "grad_norm": 0.6335748434066772,
2632
+ "learning_rate": 2.838168662762699e-06,
2633
+ "loss": 0.0133,
2634
+ "step": 375
2635
+ },
2636
+ {
2637
+ "epoch": 3.1875,
2638
+ "grad_norm": 0.5320702195167542,
2639
+ "learning_rate": 2.8252375441718137e-06,
2640
+ "loss": 0.0108,
2641
+ "step": 376
2642
+ },
2643
+ {
2644
+ "epoch": 3.196022727272727,
2645
+ "grad_norm": 0.49305084347724915,
2646
+ "learning_rate": 2.812297568177002e-06,
2647
+ "loss": 0.0107,
2648
+ "step": 377
2649
+ },
2650
+ {
2651
+ "epoch": 3.2045454545454546,
2652
+ "grad_norm": 0.5496407747268677,
2653
+ "learning_rate": 2.7993490871809808e-06,
2654
+ "loss": 0.0109,
2655
+ "step": 378
2656
+ },
2657
+ {
2658
+ "epoch": 3.2130681818181817,
2659
+ "grad_norm": 0.4906885623931885,
2660
+ "learning_rate": 2.7863924538180904e-06,
2661
+ "loss": 0.0123,
2662
+ "step": 379
2663
+ },
2664
+ {
2665
+ "epoch": 3.221590909090909,
2666
+ "grad_norm": 0.5412434339523315,
2667
+ "learning_rate": 2.773428020944687e-06,
2668
+ "loss": 0.0128,
2669
+ "step": 380
2670
+ },
2671
+ {
2672
+ "epoch": 3.2301136363636362,
2673
+ "grad_norm": 0.6251363754272461,
2674
+ "learning_rate": 2.7604561416295366e-06,
2675
+ "loss": 0.0117,
2676
+ "step": 381
2677
+ },
2678
+ {
2679
+ "epoch": 3.2386363636363638,
2680
+ "grad_norm": 0.517668604850769,
2681
+ "learning_rate": 2.747477169144202e-06,
2682
+ "loss": 0.0113,
2683
+ "step": 382
2684
+ },
2685
+ {
2686
+ "epoch": 3.247159090909091,
2687
+ "grad_norm": 0.5211403965950012,
2688
+ "learning_rate": 2.7344914569534154e-06,
2689
+ "loss": 0.0103,
2690
+ "step": 383
2691
+ },
2692
+ {
2693
+ "epoch": 3.2556818181818183,
2694
+ "grad_norm": 0.797484278678894,
2695
+ "learning_rate": 2.721499358705458e-06,
2696
+ "loss": 0.0111,
2697
+ "step": 384
2698
+ },
2699
+ {
2700
+ "epoch": 3.2642045454545454,
2701
+ "grad_norm": 0.6931593418121338,
2702
+ "learning_rate": 2.708501228222523e-06,
2703
+ "loss": 0.0115,
2704
+ "step": 385
2705
+ },
2706
+ {
2707
+ "epoch": 3.2727272727272725,
2708
+ "grad_norm": 0.515566349029541,
2709
+ "learning_rate": 2.695497419491089e-06,
2710
+ "loss": 0.0114,
2711
+ "step": 386
2712
+ },
2713
+ {
2714
+ "epoch": 3.28125,
2715
+ "grad_norm": 0.48849251866340637,
2716
+ "learning_rate": 2.682488286652269e-06,
2717
+ "loss": 0.01,
2718
+ "step": 387
2719
+ },
2720
+ {
2721
+ "epoch": 3.2897727272727275,
2722
+ "grad_norm": 0.6838280558586121,
2723
+ "learning_rate": 2.6694741839921734e-06,
2724
+ "loss": 0.011,
2725
+ "step": 388
2726
+ },
2727
+ {
2728
+ "epoch": 3.2982954545454546,
2729
+ "grad_norm": 0.4109753966331482,
2730
+ "learning_rate": 2.656455465932259e-06,
2731
+ "loss": 0.007,
2732
+ "step": 389
2733
+ },
2734
+ {
2735
+ "epoch": 3.3068181818181817,
2736
+ "grad_norm": 0.4408889412879944,
2737
+ "learning_rate": 2.6434324870196746e-06,
2738
+ "loss": 0.0096,
2739
+ "step": 390
2740
+ },
2741
+ {
2742
+ "epoch": 3.315340909090909,
2743
+ "grad_norm": 0.4267784059047699,
2744
+ "learning_rate": 2.63040560191761e-06,
2745
+ "loss": 0.009,
2746
+ "step": 391
2747
+ },
2748
+ {
2749
+ "epoch": 3.3238636363636362,
2750
+ "grad_norm": 0.42103955149650574,
2751
+ "learning_rate": 2.617375165395634e-06,
2752
+ "loss": 0.0084,
2753
+ "step": 392
2754
+ },
2755
+ {
2756
+ "epoch": 3.3323863636363638,
2757
+ "grad_norm": 0.4938065707683563,
2758
+ "learning_rate": 2.6043415323200333e-06,
2759
+ "loss": 0.0084,
2760
+ "step": 393
2761
+ },
2762
+ {
2763
+ "epoch": 3.340909090909091,
2764
+ "grad_norm": 0.9540116190910339,
2765
+ "learning_rate": 2.591305057644148e-06,
2766
+ "loss": 0.0123,
2767
+ "step": 394
2768
+ },
2769
+ {
2770
+ "epoch": 3.3494318181818183,
2771
+ "grad_norm": 0.5479457974433899,
2772
+ "learning_rate": 2.5782660963987054e-06,
2773
+ "loss": 0.01,
2774
+ "step": 395
2775
+ },
2776
+ {
2777
+ "epoch": 3.3579545454545454,
2778
+ "grad_norm": 0.44472524523735046,
2779
+ "learning_rate": 2.5652250036821522e-06,
2780
+ "loss": 0.0066,
2781
+ "step": 396
2782
+ },
2783
+ {
2784
+ "epoch": 3.3664772727272725,
2785
+ "grad_norm": 1.0710281133651733,
2786
+ "learning_rate": 2.552182134650982e-06,
2787
+ "loss": 0.0109,
2788
+ "step": 397
2789
+ },
2790
+ {
2791
+ "epoch": 3.375,
2792
+ "grad_norm": 0.6171219944953918,
2793
+ "learning_rate": 2.5391378445100646e-06,
2794
+ "loss": 0.0054,
2795
+ "step": 398
2796
+ },
2797
+ {
2798
+ "epoch": 3.3835227272727275,
2799
+ "grad_norm": 0.5634967684745789,
2800
+ "learning_rate": 2.526092488502971e-06,
2801
+ "loss": 0.0091,
2802
+ "step": 399
2803
+ },
2804
+ {
2805
+ "epoch": 3.3920454545454546,
2806
+ "grad_norm": 0.5629643797874451,
2807
+ "learning_rate": 2.5130464219022994e-06,
2808
+ "loss": 0.0079,
2809
+ "step": 400
2810
+ },
2811
+ {
2812
+ "epoch": 3.4005681818181817,
2813
+ "grad_norm": 0.45441868901252747,
2814
+ "learning_rate": 2.5e-06,
2815
+ "loss": 0.0084,
2816
+ "step": 401
2817
+ },
2818
+ {
2819
+ "epoch": 3.409090909090909,
2820
+ "grad_norm": 0.7561835646629333,
2821
+ "learning_rate": 2.4869535780977023e-06,
2822
+ "loss": 0.0148,
2823
+ "step": 402
2824
+ },
2825
+ {
2826
+ "epoch": 3.4176136363636362,
2827
+ "grad_norm": 0.42049774527549744,
2828
+ "learning_rate": 2.47390751149703e-06,
2829
+ "loss": 0.0071,
2830
+ "step": 403
2831
+ },
2832
+ {
2833
+ "epoch": 3.4261363636363638,
2834
+ "grad_norm": 0.4444538652896881,
2835
+ "learning_rate": 2.460862155489936e-06,
2836
+ "loss": 0.0075,
2837
+ "step": 404
2838
+ },
2839
+ {
2840
+ "epoch": 3.434659090909091,
2841
+ "grad_norm": 0.6667734980583191,
2842
+ "learning_rate": 2.447817865349018e-06,
2843
+ "loss": 0.0098,
2844
+ "step": 405
2845
+ },
2846
+ {
2847
+ "epoch": 3.4431818181818183,
2848
+ "grad_norm": 0.6603801250457764,
2849
+ "learning_rate": 2.4347749963178486e-06,
2850
+ "loss": 0.0079,
2851
+ "step": 406
2852
+ },
2853
+ {
2854
+ "epoch": 3.4517045454545454,
2855
+ "grad_norm": 0.5178464651107788,
2856
+ "learning_rate": 2.421733903601296e-06,
2857
+ "loss": 0.0065,
2858
+ "step": 407
2859
+ },
2860
+ {
2861
+ "epoch": 3.4602272727272725,
2862
+ "grad_norm": 0.47047677636146545,
2863
+ "learning_rate": 2.408694942355853e-06,
2864
+ "loss": 0.0084,
2865
+ "step": 408
2866
+ },
2867
+ {
2868
+ "epoch": 3.46875,
2869
+ "grad_norm": 0.4323576092720032,
2870
+ "learning_rate": 2.3956584676799676e-06,
2871
+ "loss": 0.0068,
2872
+ "step": 409
2873
+ },
2874
+ {
2875
+ "epoch": 3.4772727272727275,
2876
+ "grad_norm": 0.45331141352653503,
2877
+ "learning_rate": 2.3826248346043664e-06,
2878
+ "loss": 0.008,
2879
+ "step": 410
2880
+ },
2881
+ {
2882
+ "epoch": 3.4857954545454546,
2883
+ "grad_norm": 0.4945763051509857,
2884
+ "learning_rate": 2.369594398082391e-06,
2885
+ "loss": 0.0088,
2886
+ "step": 411
2887
+ },
2888
+ {
2889
+ "epoch": 3.4943181818181817,
2890
+ "grad_norm": 0.4799608588218689,
2891
+ "learning_rate": 2.356567512980326e-06,
2892
+ "loss": 0.0046,
2893
+ "step": 412
2894
+ },
2895
+ {
2896
+ "epoch": 3.502840909090909,
2897
+ "grad_norm": 0.5917950868606567,
2898
+ "learning_rate": 2.343544534067742e-06,
2899
+ "loss": 0.0064,
2900
+ "step": 413
2901
+ },
2902
+ {
2903
+ "epoch": 3.5113636363636362,
2904
+ "grad_norm": 0.5370076894760132,
2905
+ "learning_rate": 2.3305258160078274e-06,
2906
+ "loss": 0.0061,
2907
+ "step": 414
2908
+ },
2909
+ {
2910
+ "epoch": 3.5198863636363638,
2911
+ "grad_norm": 0.5539790391921997,
2912
+ "learning_rate": 2.317511713347731e-06,
2913
+ "loss": 0.0057,
2914
+ "step": 415
2915
+ },
2916
+ {
2917
+ "epoch": 3.528409090909091,
2918
+ "grad_norm": 0.5244786739349365,
2919
+ "learning_rate": 2.304502580508912e-06,
2920
+ "loss": 0.0077,
2921
+ "step": 416
2922
+ },
2923
+ {
2924
+ "epoch": 3.5369318181818183,
2925
+ "grad_norm": 0.5492494106292725,
2926
+ "learning_rate": 2.291498771777478e-06,
2927
+ "loss": 0.008,
2928
+ "step": 417
2929
+ },
2930
+ {
2931
+ "epoch": 3.5454545454545454,
2932
+ "grad_norm": 0.5254555940628052,
2933
+ "learning_rate": 2.278500641294543e-06,
2934
+ "loss": 0.007,
2935
+ "step": 418
2936
+ },
2937
+ {
2938
+ "epoch": 3.5539772727272725,
2939
+ "grad_norm": 0.7219369411468506,
2940
+ "learning_rate": 2.2655085430465855e-06,
2941
+ "loss": 0.0055,
2942
+ "step": 419
2943
+ },
2944
+ {
2945
+ "epoch": 3.5625,
2946
+ "grad_norm": 0.8314915895462036,
2947
+ "learning_rate": 2.252522830855798e-06,
2948
+ "loss": 0.0068,
2949
+ "step": 420
2950
+ },
2951
+ {
2952
+ "epoch": 3.5710227272727275,
2953
+ "grad_norm": 0.4091542363166809,
2954
+ "learning_rate": 2.239543858370464e-06,
2955
+ "loss": 0.004,
2956
+ "step": 421
2957
+ },
2958
+ {
2959
+ "epoch": 3.5795454545454546,
2960
+ "grad_norm": 0.6792344450950623,
2961
+ "learning_rate": 2.2265719790553147e-06,
2962
+ "loss": 0.0059,
2963
+ "step": 422
2964
+ },
2965
+ {
2966
+ "epoch": 3.5880681818181817,
2967
+ "grad_norm": 0.4202214777469635,
2968
+ "learning_rate": 2.2136075461819104e-06,
2969
+ "loss": 0.0077,
2970
+ "step": 423
2971
+ },
2972
+ {
2973
+ "epoch": 3.596590909090909,
2974
+ "grad_norm": 0.6968201398849487,
2975
+ "learning_rate": 2.2006509128190196e-06,
2976
+ "loss": 0.0086,
2977
+ "step": 424
2978
+ },
2979
+ {
2980
+ "epoch": 3.6051136363636362,
2981
+ "grad_norm": 0.6859596371650696,
2982
+ "learning_rate": 2.1877024318229987e-06,
2983
+ "loss": 0.0055,
2984
+ "step": 425
2985
+ },
2986
+ {
2987
+ "epoch": 3.6136363636363638,
2988
+ "grad_norm": 0.5254549980163574,
2989
+ "learning_rate": 2.1747624558281867e-06,
2990
+ "loss": 0.0076,
2991
+ "step": 426
2992
+ },
2993
+ {
2994
+ "epoch": 3.622159090909091,
2995
+ "grad_norm": 0.38539767265319824,
2996
+ "learning_rate": 2.1618313372373016e-06,
2997
+ "loss": 0.0058,
2998
+ "step": 427
2999
+ },
3000
+ {
3001
+ "epoch": 3.6306818181818183,
3002
+ "grad_norm": 0.5018851161003113,
3003
+ "learning_rate": 2.1489094282118393e-06,
3004
+ "loss": 0.0063,
3005
+ "step": 428
3006
+ },
3007
+ {
3008
+ "epoch": 3.6392045454545454,
3009
+ "grad_norm": 0.2640945017337799,
3010
+ "learning_rate": 2.1359970806624886e-06,
3011
+ "loss": 0.0033,
3012
+ "step": 429
3013
+ },
3014
+ {
3015
+ "epoch": 3.6477272727272725,
3016
+ "grad_norm": 0.5288704037666321,
3017
+ "learning_rate": 2.1230946462395412e-06,
3018
+ "loss": 0.0081,
3019
+ "step": 430
3020
+ },
3021
+ {
3022
+ "epoch": 3.65625,
3023
+ "grad_norm": 0.5245815515518188,
3024
+ "learning_rate": 2.1102024763233224e-06,
3025
+ "loss": 0.0053,
3026
+ "step": 431
3027
+ },
3028
+ {
3029
+ "epoch": 3.6647727272727275,
3030
+ "grad_norm": 0.4336767792701721,
3031
+ "learning_rate": 2.0973209220146135e-06,
3032
+ "loss": 0.0057,
3033
+ "step": 432
3034
+ },
3035
+ {
3036
+ "epoch": 3.6732954545454546,
3037
+ "grad_norm": 0.6342284679412842,
3038
+ "learning_rate": 2.084450334125096e-06,
3039
+ "loss": 0.0066,
3040
+ "step": 433
3041
+ },
3042
+ {
3043
+ "epoch": 3.6818181818181817,
3044
+ "grad_norm": 0.44606292247772217,
3045
+ "learning_rate": 2.071591063167797e-06,
3046
+ "loss": 0.0067,
3047
+ "step": 434
3048
+ },
3049
+ {
3050
+ "epoch": 3.690340909090909,
3051
+ "grad_norm": 0.519180953502655,
3052
+ "learning_rate": 2.0587434593475385e-06,
3053
+ "loss": 0.0046,
3054
+ "step": 435
3055
+ },
3056
+ {
3057
+ "epoch": 3.6988636363636362,
3058
+ "grad_norm": 0.44359397888183594,
3059
+ "learning_rate": 2.045907872551409e-06,
3060
+ "loss": 0.0052,
3061
+ "step": 436
3062
+ },
3063
+ {
3064
+ "epoch": 3.7073863636363638,
3065
+ "grad_norm": 0.43811115622520447,
3066
+ "learning_rate": 2.033084652339226e-06,
3067
+ "loss": 0.0048,
3068
+ "step": 437
3069
+ },
3070
+ {
3071
+ "epoch": 3.715909090909091,
3072
+ "grad_norm": 0.5960409045219421,
3073
+ "learning_rate": 2.0202741479340193e-06,
3074
+ "loss": 0.0063,
3075
+ "step": 438
3076
+ },
3077
+ {
3078
+ "epoch": 3.7244318181818183,
3079
+ "grad_norm": 0.4575386345386505,
3080
+ "learning_rate": 2.0074767082125244e-06,
3081
+ "loss": 0.0052,
3082
+ "step": 439
3083
+ },
3084
+ {
3085
+ "epoch": 3.7329545454545454,
3086
+ "grad_norm": 0.6704768538475037,
3087
+ "learning_rate": 1.9946926816956743e-06,
3088
+ "loss": 0.0075,
3089
+ "step": 440
3090
+ },
3091
+ {
3092
+ "epoch": 3.7414772727272725,
3093
+ "grad_norm": 0.44066664576530457,
3094
+ "learning_rate": 1.9819224165391145e-06,
3095
+ "loss": 0.004,
3096
+ "step": 441
3097
+ },
3098
+ {
3099
+ "epoch": 3.75,
3100
+ "grad_norm": 0.5848504304885864,
3101
+ "learning_rate": 1.969166260523717e-06,
3102
+ "loss": 0.005,
3103
+ "step": 442
3104
+ },
3105
+ {
3106
+ "epoch": 3.7585227272727275,
3107
+ "grad_norm": 0.43681785464286804,
3108
+ "learning_rate": 1.956424561046108e-06,
3109
+ "loss": 0.0041,
3110
+ "step": 443
3111
+ },
3112
+ {
3113
+ "epoch": 3.7670454545454546,
3114
+ "grad_norm": 0.31628116965293884,
3115
+ "learning_rate": 1.9436976651092143e-06,
3116
+ "loss": 0.0026,
3117
+ "step": 444
3118
+ },
3119
+ {
3120
+ "epoch": 3.7755681818181817,
3121
+ "grad_norm": 0.4530532658100128,
3122
+ "learning_rate": 1.9309859193128055e-06,
3123
+ "loss": 0.0052,
3124
+ "step": 445
3125
+ },
3126
+ {
3127
+ "epoch": 3.784090909090909,
3128
+ "grad_norm": 0.258119136095047,
3129
+ "learning_rate": 1.918289669844058e-06,
3130
+ "loss": 0.0025,
3131
+ "step": 446
3132
+ },
3133
+ {
3134
+ "epoch": 3.7926136363636362,
3135
+ "grad_norm": 0.40686145424842834,
3136
+ "learning_rate": 1.905609262468128e-06,
3137
+ "loss": 0.005,
3138
+ "step": 447
3139
+ },
3140
+ {
3141
+ "epoch": 3.8011363636363638,
3142
+ "grad_norm": 0.4893140196800232,
3143
+ "learning_rate": 1.8929450425187298e-06,
3144
+ "loss": 0.0033,
3145
+ "step": 448
3146
+ },
3147
+ {
3148
+ "epoch": 3.809659090909091,
3149
+ "grad_norm": 0.8415848612785339,
3150
+ "learning_rate": 1.880297354888741e-06,
3151
+ "loss": 0.0044,
3152
+ "step": 449
3153
+ },
3154
+ {
3155
+ "epoch": 3.8181818181818183,
3156
+ "grad_norm": 0.21848903596401215,
3157
+ "learning_rate": 1.8676665440207982e-06,
3158
+ "loss": 0.0015,
3159
+ "step": 450
3160
+ },
3161
+ {
3162
+ "epoch": 3.8267045454545454,
3163
+ "grad_norm": 0.7620615363121033,
3164
+ "learning_rate": 1.8550529538979273e-06,
3165
+ "loss": 0.0069,
3166
+ "step": 451
3167
+ },
3168
+ {
3169
+ "epoch": 3.8352272727272725,
3170
+ "grad_norm": 0.35021477937698364,
3171
+ "learning_rate": 1.8424569280341653e-06,
3172
+ "loss": 0.004,
3173
+ "step": 452
3174
+ },
3175
+ {
3176
+ "epoch": 3.84375,
3177
+ "grad_norm": 0.477241575717926,
3178
+ "learning_rate": 1.8298788094652158e-06,
3179
+ "loss": 0.0052,
3180
+ "step": 453
3181
+ },
3182
+ {
3183
+ "epoch": 3.8522727272727275,
3184
+ "grad_norm": 0.3984035849571228,
3185
+ "learning_rate": 1.817318940739098e-06,
3186
+ "loss": 0.0029,
3187
+ "step": 454
3188
+ },
3189
+ {
3190
+ "epoch": 3.8607954545454546,
3191
+ "grad_norm": 0.3424110412597656,
3192
+ "learning_rate": 1.8047776639068204e-06,
3193
+ "loss": 0.0044,
3194
+ "step": 455
3195
+ },
3196
+ {
3197
+ "epoch": 3.8693181818181817,
3198
+ "grad_norm": 1.0182089805603027,
3199
+ "learning_rate": 1.7922553205130708e-06,
3200
+ "loss": 0.0027,
3201
+ "step": 456
3202
+ },
3203
+ {
3204
+ "epoch": 3.877840909090909,
3205
+ "grad_norm": 0.4957254230976105,
3206
+ "learning_rate": 1.7797522515869062e-06,
3207
+ "loss": 0.0028,
3208
+ "step": 457
3209
+ },
3210
+ {
3211
+ "epoch": 3.8863636363636362,
3212
+ "grad_norm": 0.42411452531814575,
3213
+ "learning_rate": 1.767268797632472e-06,
3214
+ "loss": 0.0032,
3215
+ "step": 458
3216
+ },
3217
+ {
3218
+ "epoch": 3.8948863636363638,
3219
+ "grad_norm": 0.3485550582408905,
3220
+ "learning_rate": 1.7548052986197258e-06,
3221
+ "loss": 0.0032,
3222
+ "step": 459
3223
+ },
3224
+ {
3225
+ "epoch": 3.903409090909091,
3226
+ "grad_norm": 0.3125132918357849,
3227
+ "learning_rate": 1.7423620939751787e-06,
3228
+ "loss": 0.0017,
3229
+ "step": 460
3230
+ },
3231
+ {
3232
+ "epoch": 3.9119318181818183,
3233
+ "grad_norm": 0.7568768262863159,
3234
+ "learning_rate": 1.7299395225726533e-06,
3235
+ "loss": 0.006,
3236
+ "step": 461
3237
+ },
3238
+ {
3239
+ "epoch": 3.9204545454545454,
3240
+ "grad_norm": 0.7975472807884216,
3241
+ "learning_rate": 1.7175379227240524e-06,
3242
+ "loss": 0.0044,
3243
+ "step": 462
3244
+ },
3245
+ {
3246
+ "epoch": 3.9289772727272725,
3247
+ "grad_norm": 0.5542371273040771,
3248
+ "learning_rate": 1.7051576321701493e-06,
3249
+ "loss": 0.0034,
3250
+ "step": 463
3251
+ },
3252
+ {
3253
+ "epoch": 3.9375,
3254
+ "grad_norm": 0.3784295916557312,
3255
+ "learning_rate": 1.6927989880713852e-06,
3256
+ "loss": 0.0024,
3257
+ "step": 464
3258
+ },
3259
+ {
3260
+ "epoch": 3.9460227272727275,
3261
+ "grad_norm": 0.7081852555274963,
3262
+ "learning_rate": 1.680462326998689e-06,
3263
+ "loss": 0.0047,
3264
+ "step": 465
3265
+ },
3266
+ {
3267
+ "epoch": 3.9545454545454546,
3268
+ "grad_norm": 0.268210232257843,
3269
+ "learning_rate": 1.6681479849243153e-06,
3270
+ "loss": 0.0019,
3271
+ "step": 466
3272
+ },
3273
+ {
3274
+ "epoch": 3.9630681818181817,
3275
+ "grad_norm": 0.8834506273269653,
3276
+ "learning_rate": 1.6558562972126869e-06,
3277
+ "loss": 0.0025,
3278
+ "step": 467
3279
+ },
3280
+ {
3281
+ "epoch": 3.971590909090909,
3282
+ "grad_norm": 0.8946535587310791,
3283
+ "learning_rate": 1.6435875986112685e-06,
3284
+ "loss": 0.0035,
3285
+ "step": 468
3286
+ }
3287
+ ],
3288
+ "logging_steps": 1,
3289
+ "max_steps": 702,
3290
+ "num_input_tokens_seen": 0,
3291
+ "num_train_epochs": 6,
3292
+ "save_steps": 117,
3293
+ "stateful_callbacks": {
3294
+ "TrainerControl": {
3295
+ "args": {
3296
+ "should_epoch_stop": false,
3297
+ "should_evaluate": false,
3298
+ "should_log": false,
3299
+ "should_save": true,
3300
+ "should_training_stop": false
3301
+ },
3302
+ "attributes": {}
3303
+ }
3304
+ },
3305
+ "total_flos": 1.0433466041169347e+18,
3306
+ "train_batch_size": 4,
3307
+ "trial_name": null,
3308
+ "trial_params": null
3309
+ }
checkpoint-468/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52792c1a622772d0c05d1c4274260ef25199afed15af6dc0e0643f0449bb5ce6
3
+ size 8056
checkpoint-468/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-468/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)