Hugging Face's logo

RayDu0010
/
tulu10k_klora

Safetensors
Model card Files
xet
 Community
RayDu0010 commited on
Commit
aee1667
·
verified ·
1 Parent(s): 36194fe

Upload folder using huggingface_hub

Browse files
Files changed (14) hide show
  1. instruct/127_128_e3_3e-5/adapter_config.json +39 -0
  2. instruct/127_128_e3_3e-5/adapter_model.safetensors +3 -0
  3. instruct/127_128_e3_3e-5/added_tokens.json +9 -0
  4. instruct/127_128_e3_3e-5/all_results.json +9 -0
  5. instruct/127_128_e3_3e-5/chat_template.jinja +62 -0
  6. instruct/127_128_e3_3e-5/config.json +32 -0
  7. instruct/127_128_e3_3e-5/merges.txt +0 -0
  8. instruct/127_128_e3_3e-5/special_tokens_map.json +33 -0
  9. instruct/127_128_e3_3e-5/tokenizer.json +0 -0
  10. instruct/127_128_e3_3e-5/tokenizer_config.json +234 -0
  11. instruct/127_128_e3_3e-5/train_results.json +9 -0
  12. instruct/127_128_e3_3e-5/trainer_state.json +3109 -0
  13. instruct/127_128_e3_3e-5/training_args.bin +3 -0
  14. instruct/127_128_e3_3e-5/vocab.json +0 -0
instruct/127_128_e3_3e-5/adapter_config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "ibm-granite/granite-3.3-8b-instruct",
5
+ "bias": "none",
6
+ "corda_config": null,
7
+ "eva_config": null,
8
+ "exclude_modules": null,
9
+ "fan_in_fan_out": false,
10
+ "inference_mode": true,
11
+ "init_lora_weights": true,
12
+ "layer_replication": null,
13
+ "layers_pattern": null,
14
+ "layers_to_transform": null,
15
+ "loftq_config": {},
16
+ "lora_alpha": 256,
17
+ "lora_bias": false,
18
+ "lora_dropout": 0.05,
19
+ "megatron_config": null,
20
+ "megatron_core": "megatron.core",
21
+ "modules_to_save": null,
22
+ "peft_type": "LORA",
23
+ "r": 128,
24
+ "rank_pattern": {},
25
+ "revision": null,
26
+ "target_modules": [
27
+ "v_proj",
28
+ "up_proj",
29
+ "q_proj",
30
+ "o_proj",
31
+ "gate_proj",
32
+ "down_proj",
33
+ "k_proj"
34
+ ],
35
+ "task_type": "CAUSAL_LM",
36
+ "trainable_token_indices": null,
37
+ "use_dora": false,
38
+ "use_rslora": false
39
+ }
instruct/127_128_e3_3e-5/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15d536a13f0686dc7a7a386681233514aabc8c97d24b5447f71af175c4a6633d
3
+ size 791751704
instruct/127_128_e3_3e-5/added_tokens.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "<|end_of_cite|>": 49156,
3
+ "<|end_of_plugin|>": 49158,
4
+ "<|end_of_role|>": 49153,
5
+ "<|start_of_cite|>": 49155,
6
+ "<|start_of_plugin|>": 49157,
7
+ "<|start_of_role|>": 49152,
8
+ "<|tool_call|>": 49154
9
+ }
instruct/127_128_e3_3e-5/all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 3.0054141731243295e+18,
4
+ "train_loss": 0.341107062958062,
5
+ "train_runtime": 1404.3398,
6
+ "train_samples": 23386,
7
+ "train_samples_per_second": 49.958,
8
+ "train_steps_per_second": 1.562
9
+ }
instruct/127_128_e3_3e-5/chat_template.jinja ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {# Alias tools -> available_tools #}
2
+ {%- if tools and not available_tools -%}
3
+ {%- set available_tools = tools -%}
4
+ {%- endif -%}
5
+ {%- if messages[0]['role'] == 'system' %}
6
+ {%- set system_message = messages[0]['content'] %}
7
+ {%- set loop_messages = messages[1:] %}
8
+ {%- else %}
9
+ {%- set system_message = "Knowledge Cutoff Date: April 2024.
10
+ Today's Date: " + strftime_now('%B %d, %Y') + ".
11
+ You are Granite, developed by IBM." %}
12
+ {%- if available_tools and documents %}
13
+ {%- set system_message = system_message + " You are a helpful assistant with access to the following tools. When a tool is required to answer the user's query, respond only with <|tool_call|> followed by a JSON list of tools used. If a tool does not exist in the provided list of tools, notify the user that you do not have the ability to fulfill the request.
14
+ Write the response to the user's input by strictly aligning with the facts in the provided documents. If the information needed to answer the question is not available in the documents, inform the user that the question cannot be answered based on the available data." %}
15
+ {%- elif available_tools %}
16
+ {%- set system_message = system_message + " You are a helpful assistant with access to the following tools. When a tool is required to answer the user's query, respond only with <|tool_call|> followed by a JSON list of tools used. If a tool does not exist in the provided list of tools, notify the user that you do not have the ability to fulfill the request." %}
17
+ {%- elif documents %}
18
+ {%- set system_message = system_message + " Write the response to the user's input by strictly aligning with the facts in the provided documents. If the information needed to answer the question is not available in the documents, inform the user that the question cannot be answered based on the available data." %}
19
+ {%- elif thinking %}
20
+ {%- set system_message = system_message + " You are a helpful AI assistant.
21
+ Respond to every user query in a comprehensive and detailed way. You can write down your thoughts and reasoning process before responding. In the thought process, engage in a comprehensive cycle of analysis, summarization, exploration, reassessment, reflection, backtracing, and iteration to develop well-considered thinking process. In the response section, based on various attempts, explorations, and reflections from the thoughts section, systematically present the final solution that you deem correct. The response should summarize the thought process. Write your thoughts between <think></think> and write your response between <response></response> for each user query." %}
22
+ {%- else %}
23
+ {%- set system_message = system_message + " You are a helpful AI assistant." %}
24
+ {%- endif %}
25
+ {%- if 'citations' in controls and documents %}
26
+ {%- set system_message = system_message + '
27
+ Use the symbols <|start_of_cite|> and <|end_of_cite|> to indicate when a fact comes from a document in the search result, e.g <|start_of_cite|> {document_id: 1}my fact <|end_of_cite|> for a fact from document 1. Afterwards, list all the citations with their corresponding documents in an ordered list.' %}
28
+ {%- endif %}
29
+ {%- if 'hallucinations' in controls and documents %}
30
+ {%- set system_message = system_message + '
31
+ Finally, after the response is written, include a numbered list of sentences from the response with a corresponding risk value that are hallucinated and not based in the documents.' %}
32
+ {%- endif %}
33
+ {%- set loop_messages = messages %}
34
+ {%- endif %}
35
+ {{- '<|start_of_role|>system<|end_of_role|>' + system_message + '<|end_of_text|>
36
+ ' }}
37
+ {%- if available_tools %}
38
+ {{- '<|start_of_role|>available_tools<|end_of_role|>' }}
39
+ {{- available_tools | tojson(indent=4) }}
40
+ {{- '<|end_of_text|>
41
+ ' }}
42
+ {%- endif %}
43
+ {%- if documents %}
44
+ {%- for document in documents %}
45
+ {{- '<|start_of_role|>document {"document_id": "' + document['doc_id'] | string + '"}<|end_of_role|>
46
+ ' }}
47
+ {{- document['text'] }}
48
+ {{- '<|end_of_text|>
49
+ ' }}
50
+ {%- endfor %}
51
+ {%- endif %}
52
+ {%- for message in loop_messages %}
53
+ {{- '<|start_of_role|>' + message['role'] + '<|end_of_role|>' + message['content'] + '<|end_of_text|>
54
+ ' }}
55
+ {%- if loop.last and add_generation_prompt %}
56
+ {{- '<|start_of_role|>assistant' }}
57
+ {%- if controls %}
58
+ {{- ' ' + controls | tojson()}}
59
+ {%- endif %}
60
+ {{- '<|end_of_role|>' }}
61
+ {%- endif %}
62
+ {%- endfor %}
instruct/127_128_e3_3e-5/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "GraniteForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "attention_multiplier": 0.0078125,
8
+ "bos_token_id": 0,
9
+ "embedding_multiplier": 12.0,
10
+ "eos_token_id": 0,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 4096,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 12800,
15
+ "logits_scaling": 16.0,
16
+ "max_position_embeddings": 131072,
17
+ "mlp_bias": false,
18
+ "model_type": "granite",
19
+ "num_attention_heads": 32,
20
+ "num_hidden_layers": 40,
21
+ "num_key_value_heads": 8,
22
+ "pad_token_id": 0,
23
+ "residual_multiplier": 0.22,
24
+ "rms_norm_eps": 1e-05,
25
+ "rope_scaling": null,
26
+ "rope_theta": 10000000.0,
27
+ "tie_word_embeddings": true,
28
+ "torch_dtype": "bfloat16",
29
+ "transformers_version": "4.52.4",
30
+ "use_cache": true,
31
+ "vocab_size": 49159
32
+ }
instruct/127_128_e3_3e-5/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
instruct/127_128_e3_3e-5/special_tokens_map.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|start_of_role|>",
4
+ "<|end_of_role|>",
5
+ "<|tool_call|>",
6
+ "<|start_of_cite|>",
7
+ "<|end_of_cite|>",
8
+ "<|start_of_plugin|>",
9
+ "<|end_of_plugin|>"
10
+ ],
11
+ "bos_token": {
12
+ "content": "<|end_of_text|>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "eos_token": {
19
+ "content": "<|end_of_text|>",
20
+ "lstrip": false,
21
+ "normalized": false,
22
+ "rstrip": false,
23
+ "single_word": false
24
+ },
25
+ "pad_token": "<|end_of_plugin|>",
26
+ "unk_token": {
27
+ "content": "<|end_of_text|>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ }
33
+ }
instruct/127_128_e3_3e-5/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
instruct/127_128_e3_3e-5/tokenizer_config.json ADDED
@@ -0,0 +1,234 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<|end_of_text|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<fim_prefix>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "<fim_middle>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "3": {
30
+ "content": "<fim_suffix>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "4": {
38
+ "content": "<fim_pad>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "5": {
46
+ "content": "<filename>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "6": {
54
+ "content": "<gh_stars>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "7": {
62
+ "content": "<issue_start>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "8": {
70
+ "content": "<issue_comment>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "9": {
78
+ "content": "<issue_closed>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "10": {
86
+ "content": "<jupyter_start>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "11": {
94
+ "content": "<jupyter_text>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "12": {
102
+ "content": "<jupyter_code>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "13": {
110
+ "content": "<jupyter_output>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "14": {
118
+ "content": "<empty_output>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": true
124
+ },
125
+ "15": {
126
+ "content": "<commit_before>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": true
132
+ },
133
+ "16": {
134
+ "content": "<commit_msg>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": true
140
+ },
141
+ "17": {
142
+ "content": "<commit_after>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": true
148
+ },
149
+ "18": {
150
+ "content": "<reponame>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": true
156
+ },
157
+ "49152": {
158
+ "content": "<|start_of_role|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": true
164
+ },
165
+ "49153": {
166
+ "content": "<|end_of_role|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": true
172
+ },
173
+ "49154": {
174
+ "content": "<|tool_call|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": true
180
+ },
181
+ "49155": {
182
+ "content": "<|start_of_cite|>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": true
188
+ },
189
+ "49156": {
190
+ "content": "<|end_of_cite|>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": true
196
+ },
197
+ "49157": {
198
+ "content": "<|start_of_plugin|>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": true
204
+ },
205
+ "49158": {
206
+ "content": "<|end_of_plugin|>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": true
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
+ "<|start_of_role|>",
216
+ "<|end_of_role|>",
217
+ "<|tool_call|>",
218
+ "<|start_of_cite|>",
219
+ "<|end_of_cite|>",
220
+ "<|start_of_plugin|>",
221
+ "<|end_of_plugin|>"
222
+ ],
223
+ "bos_token": "<|end_of_text|>",
224
+ "clean_up_tokenization_spaces": true,
225
+ "eos_token": "<|end_of_text|>",
226
+ "errors": "replace",
227
+ "extra_special_tokens": {},
228
+ "model_max_length": 8192,
229
+ "pad_token": "<|end_of_plugin|>",
230
+ "padding_side": "left",
231
+ "tokenizer_class": "GPT2Tokenizer",
232
+ "unk_token": "<|end_of_text|>",
233
+ "vocab_size": 49152
234
+ }
instruct/127_128_e3_3e-5/train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 3.0054141731243295e+18,
4
+ "train_loss": 0.341107062958062,
5
+ "train_runtime": 1404.3398,
6
+ "train_samples": 23386,
7
+ "train_samples_per_second": 49.958,
8
+ "train_steps_per_second": 1.562
9
+ }
instruct/127_128_e3_3e-5/trainer_state.json ADDED
@@ -0,0 +1,3109 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.0,
6
+ "eval_steps": 500,
7
+ "global_step": 2193,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.006839945280437756,
14
+ "grad_norm": 2.4454569816589355,
15
+ "learning_rate": 1.090909090909091e-06,
16
+ "loss": 1.4998,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.013679890560875513,
21
+ "grad_norm": 2.1232502460479736,
22
+ "learning_rate": 2.454545454545455e-06,
23
+ "loss": 1.4775,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.02051983584131327,
28
+ "grad_norm": 1.4666121006011963,
29
+ "learning_rate": 3.818181818181818e-06,
30
+ "loss": 1.4408,
31
+ "step": 15
32
+ },
33
+ {
34
+ "epoch": 0.027359781121751026,
35
+ "grad_norm": 0.9650471210479736,
36
+ "learning_rate": 5.181818181818182e-06,
37
+ "loss": 1.4299,
38
+ "step": 20
39
+ },
40
+ {
41
+ "epoch": 0.03419972640218878,
42
+ "grad_norm": 0.5837101936340332,
43
+ "learning_rate": 6.545454545454545e-06,
44
+ "loss": 1.3775,
45
+ "step": 25
46
+ },
47
+ {
48
+ "epoch": 0.04103967168262654,
49
+ "grad_norm": 0.542555034160614,
50
+ "learning_rate": 7.909090909090909e-06,
51
+ "loss": 1.3363,
52
+ "step": 30
53
+ },
54
+ {
55
+ "epoch": 0.047879616963064295,
56
+ "grad_norm": 0.38645312190055847,
57
+ "learning_rate": 9.272727272727273e-06,
58
+ "loss": 1.3719,
59
+ "step": 35
60
+ },
61
+ {
62
+ "epoch": 0.05471956224350205,
63
+ "grad_norm": 0.39401039481163025,
64
+ "learning_rate": 1.0636363636363636e-05,
65
+ "loss": 1.3675,
66
+ "step": 40
67
+ },
68
+ {
69
+ "epoch": 0.06155950752393981,
70
+ "grad_norm": 0.38187530636787415,
71
+ "learning_rate": 1.2e-05,
72
+ "loss": 1.3307,
73
+ "step": 45
74
+ },
75
+ {
76
+ "epoch": 0.06839945280437756,
77
+ "grad_norm": 0.3517046868801117,
78
+ "learning_rate": 1.3363636363636364e-05,
79
+ "loss": 1.2977,
80
+ "step": 50
81
+ },
82
+ {
83
+ "epoch": 0.07523939808481532,
84
+ "grad_norm": 0.4090668559074402,
85
+ "learning_rate": 1.4727272727272728e-05,
86
+ "loss": 1.2889,
87
+ "step": 55
88
+ },
89
+ {
90
+ "epoch": 0.08207934336525308,
91
+ "grad_norm": 0.37684422731399536,
92
+ "learning_rate": 1.6090909090909092e-05,
93
+ "loss": 1.2774,
94
+ "step": 60
95
+ },
96
+ {
97
+ "epoch": 0.08891928864569083,
98
+ "grad_norm": 0.4667328894138336,
99
+ "learning_rate": 1.7454545454545456e-05,
100
+ "loss": 1.2958,
101
+ "step": 65
102
+ },
103
+ {
104
+ "epoch": 0.09575923392612859,
105
+ "grad_norm": 0.4185400903224945,
106
+ "learning_rate": 1.881818181818182e-05,
107
+ "loss": 1.292,
108
+ "step": 70
109
+ },
110
+ {
111
+ "epoch": 0.10259917920656635,
112
+ "grad_norm": 0.43335795402526855,
113
+ "learning_rate": 2.0181818181818183e-05,
114
+ "loss": 1.2679,
115
+ "step": 75
116
+ },
117
+ {
118
+ "epoch": 0.1094391244870041,
119
+ "grad_norm": 0.3911260962486267,
120
+ "learning_rate": 2.1545454545454544e-05,
121
+ "loss": 1.2286,
122
+ "step": 80
123
+ },
124
+ {
125
+ "epoch": 0.11627906976744186,
126
+ "grad_norm": 0.44880130887031555,
127
+ "learning_rate": 2.290909090909091e-05,
128
+ "loss": 1.2258,
129
+ "step": 85
130
+ },
131
+ {
132
+ "epoch": 0.12311901504787962,
133
+ "grad_norm": 0.42907172441482544,
134
+ "learning_rate": 2.4272727272727275e-05,
135
+ "loss": 1.2275,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.12995896032831739,
140
+ "grad_norm": 0.42201051115989685,
141
+ "learning_rate": 2.5636363636363635e-05,
142
+ "loss": 1.2343,
143
+ "step": 95
144
+ },
145
+ {
146
+ "epoch": 0.13679890560875513,
147
+ "grad_norm": 0.4677179753780365,
148
+ "learning_rate": 2.7000000000000002e-05,
149
+ "loss": 1.143,
150
+ "step": 100
151
+ },
152
+ {
153
+ "epoch": 0.1436388508891929,
154
+ "grad_norm": 0.4903288185596466,
155
+ "learning_rate": 2.8363636363636363e-05,
156
+ "loss": 1.1672,
157
+ "step": 105
158
+ },
159
+ {
160
+ "epoch": 0.15047879616963064,
161
+ "grad_norm": 0.4819750189781189,
162
+ "learning_rate": 2.972727272727273e-05,
163
+ "loss": 1.1414,
164
+ "step": 110
165
+ },
166
+ {
167
+ "epoch": 0.1573187414500684,
168
+ "grad_norm": 0.5264599323272705,
169
+ "learning_rate": 2.9999727038664486e-05,
170
+ "loss": 1.1553,
171
+ "step": 115
172
+ },
173
+ {
174
+ "epoch": 0.16415868673050615,
175
+ "grad_norm": 0.4937070608139038,
176
+ "learning_rate": 2.9998618150265143e-05,
177
+ "loss": 1.1917,
178
+ "step": 120
179
+ },
180
+ {
181
+ "epoch": 0.17099863201094392,
182
+ "grad_norm": 0.5141596794128418,
183
+ "learning_rate": 2.9996656337728994e-05,
184
+ "loss": 1.1341,
185
+ "step": 125
186
+ },
187
+ {
188
+ "epoch": 0.17783857729138167,
189
+ "grad_norm": 0.47989073395729065,
190
+ "learning_rate": 2.999384171261812e-05,
191
+ "loss": 1.101,
192
+ "step": 130
193
+ },
194
+ {
195
+ "epoch": 0.18467852257181944,
196
+ "grad_norm": 0.577491283416748,
197
+ "learning_rate": 2.999017443499139e-05,
198
+ "loss": 1.171,
199
+ "step": 135
200
+ },
201
+ {
202
+ "epoch": 0.19151846785225718,
203
+ "grad_norm": 0.5731332898139954,
204
+ "learning_rate": 2.9985654713395316e-05,
205
+ "loss": 1.0856,
206
+ "step": 140
207
+ },
208
+ {
209
+ "epoch": 0.19835841313269495,
210
+ "grad_norm": 0.5869550108909607,
211
+ "learning_rate": 2.9980282804852227e-05,
212
+ "loss": 1.1259,
213
+ "step": 145
214
+ },
215
+ {
216
+ "epoch": 0.2051983584131327,
217
+ "grad_norm": 0.5815097093582153,
218
+ "learning_rate": 2.9974059014845614e-05,
219
+ "loss": 1.1109,
220
+ "step": 150
221
+ },
222
+ {
223
+ "epoch": 0.21203830369357046,
224
+ "grad_norm": 0.58452308177948,
225
+ "learning_rate": 2.996698369730281e-05,
226
+ "loss": 1.0703,
227
+ "step": 155
228
+ },
229
+ {
230
+ "epoch": 0.2188782489740082,
231
+ "grad_norm": 0.6075032949447632,
232
+ "learning_rate": 2.995905725457482e-05,
233
+ "loss": 1.0721,
234
+ "step": 160
235
+ },
236
+ {
237
+ "epoch": 0.22571819425444598,
238
+ "grad_norm": 0.6801859140396118,
239
+ "learning_rate": 2.9950280137413452e-05,
240
+ "loss": 0.9935,
241
+ "step": 165
242
+ },
243
+ {
244
+ "epoch": 0.23255813953488372,
245
+ "grad_norm": 0.7454715371131897,
246
+ "learning_rate": 2.9940652844945697e-05,
247
+ "loss": 1.0871,
248
+ "step": 170
249
+ },
250
+ {
251
+ "epoch": 0.2393980848153215,
252
+ "grad_norm": 0.6518920063972473,
253
+ "learning_rate": 2.9930175924645328e-05,
254
+ "loss": 1.0477,
255
+ "step": 175
256
+ },
257
+ {
258
+ "epoch": 0.24623803009575923,
259
+ "grad_norm": 0.6701561212539673,
260
+ "learning_rate": 2.9918849972301778e-05,
261
+ "loss": 1.0356,
262
+ "step": 180
263
+ },
264
+ {
265
+ "epoch": 0.253077975376197,
266
+ "grad_norm": 0.7389035820960999,
267
+ "learning_rate": 2.9906675631986255e-05,
268
+ "loss": 1.0251,
269
+ "step": 185
270
+ },
271
+ {
272
+ "epoch": 0.25991792065663477,
273
+ "grad_norm": 0.6748508214950562,
274
+ "learning_rate": 2.9893653596015122e-05,
275
+ "loss": 0.9816,
276
+ "step": 190
277
+ },
278
+ {
279
+ "epoch": 0.2667578659370725,
280
+ "grad_norm": 0.7654222249984741,
281
+ "learning_rate": 2.9879784604910515e-05,
282
+ "loss": 0.9615,
283
+ "step": 195
284
+ },
285
+ {
286
+ "epoch": 0.27359781121751026,
287
+ "grad_norm": 0.7578948736190796,
288
+ "learning_rate": 2.986506944735824e-05,
289
+ "loss": 0.9602,
290
+ "step": 200
291
+ },
292
+ {
293
+ "epoch": 0.280437756497948,
294
+ "grad_norm": 0.7674322724342346,
295
+ "learning_rate": 2.9849508960162927e-05,
296
+ "loss": 0.994,
297
+ "step": 205
298
+ },
299
+ {
300
+ "epoch": 0.2872777017783858,
301
+ "grad_norm": 0.7107180953025818,
302
+ "learning_rate": 2.9833104028200432e-05,
303
+ "loss": 0.9686,
304
+ "step": 210
305
+ },
306
+ {
307
+ "epoch": 0.29411764705882354,
308
+ "grad_norm": 0.7623671889305115,
309
+ "learning_rate": 2.981585558436753e-05,
310
+ "loss": 0.9534,
311
+ "step": 215
312
+ },
313
+ {
314
+ "epoch": 0.3009575923392613,
315
+ "grad_norm": 0.7795570492744446,
316
+ "learning_rate": 2.9797764609528842e-05,
317
+ "loss": 0.9243,
318
+ "step": 220
319
+ },
320
+ {
321
+ "epoch": 0.307797537619699,
322
+ "grad_norm": 0.8374460935592651,
323
+ "learning_rate": 2.977883213246109e-05,
324
+ "loss": 0.9059,
325
+ "step": 225
326
+ },
327
+ {
328
+ "epoch": 0.3146374829001368,
329
+ "grad_norm": 0.757520318031311,
330
+ "learning_rate": 2.975905922979457e-05,
331
+ "loss": 0.9012,
332
+ "step": 230
333
+ },
334
+ {
335
+ "epoch": 0.32147742818057456,
336
+ "grad_norm": 0.8288528323173523,
337
+ "learning_rate": 2.9738447025951922e-05,
338
+ "loss": 0.9031,
339
+ "step": 235
340
+ },
341
+ {
342
+ "epoch": 0.3283173734610123,
343
+ "grad_norm": 0.8580222129821777,
344
+ "learning_rate": 2.971699669308421e-05,
345
+ "loss": 0.9279,
346
+ "step": 240
347
+ },
348
+ {
349
+ "epoch": 0.33515731874145005,
350
+ "grad_norm": 0.8569541573524475,
351
+ "learning_rate": 2.9694709451004255e-05,
352
+ "loss": 0.8587,
353
+ "step": 245
354
+ },
355
+ {
356
+ "epoch": 0.34199726402188785,
357
+ "grad_norm": 0.8050552606582642,
358
+ "learning_rate": 2.9671586567117256e-05,
359
+ "loss": 0.8625,
360
+ "step": 250
361
+ },
362
+ {
363
+ "epoch": 0.3488372093023256,
364
+ "grad_norm": 0.8597307205200195,
365
+ "learning_rate": 2.9647629356348736e-05,
366
+ "loss": 0.9248,
367
+ "step": 255
368
+ },
369
+ {
370
+ "epoch": 0.35567715458276333,
371
+ "grad_norm": 0.8131337761878967,
372
+ "learning_rate": 2.9622839181069756e-05,
373
+ "loss": 0.8292,
374
+ "step": 260
375
+ },
376
+ {
377
+ "epoch": 0.3625170998632011,
378
+ "grad_norm": 0.9674245119094849,
379
+ "learning_rate": 2.9597217451019442e-05,
380
+ "loss": 0.8687,
381
+ "step": 265
382
+ },
383
+ {
384
+ "epoch": 0.3693570451436389,
385
+ "grad_norm": 0.9852687120437622,
386
+ "learning_rate": 2.9570765623224815e-05,
387
+ "loss": 0.8286,
388
+ "step": 270
389
+ },
390
+ {
391
+ "epoch": 0.3761969904240766,
392
+ "grad_norm": 0.9519874453544617,
393
+ "learning_rate": 2.9543485201917952e-05,
394
+ "loss": 0.8904,
395
+ "step": 275
396
+ },
397
+ {
398
+ "epoch": 0.38303693570451436,
399
+ "grad_norm": 0.9735475182533264,
400
+ "learning_rate": 2.9515377738450414e-05,
401
+ "loss": 0.8406,
402
+ "step": 280
403
+ },
404
+ {
405
+ "epoch": 0.3898768809849521,
406
+ "grad_norm": 0.7995354533195496,
407
+ "learning_rate": 2.9486444831205056e-05,
408
+ "loss": 0.8054,
409
+ "step": 285
410
+ },
411
+ {
412
+ "epoch": 0.3967168262653899,
413
+ "grad_norm": 0.9693818688392639,
414
+ "learning_rate": 2.9456688125505105e-05,
415
+ "loss": 0.8134,
416
+ "step": 290
417
+ },
418
+ {
419
+ "epoch": 0.40355677154582764,
420
+ "grad_norm": 0.9003785252571106,
421
+ "learning_rate": 2.942610931352062e-05,
422
+ "loss": 0.8112,
423
+ "step": 295
424
+ },
425
+ {
426
+ "epoch": 0.4103967168262654,
427
+ "grad_norm": 1.0292534828186035,
428
+ "learning_rate": 2.939471013417226e-05,
429
+ "loss": 0.81,
430
+ "step": 300
431
+ },
432
+ {
433
+ "epoch": 0.4172366621067031,
434
+ "grad_norm": 1.007619857788086,
435
+ "learning_rate": 2.936249237303237e-05,
436
+ "loss": 0.7881,
437
+ "step": 305
438
+ },
439
+ {
440
+ "epoch": 0.4240766073871409,
441
+ "grad_norm": 0.9168869853019714,
442
+ "learning_rate": 2.9329457862223474e-05,
443
+ "loss": 0.7881,
444
+ "step": 310
445
+ },
446
+ {
447
+ "epoch": 0.43091655266757867,
448
+ "grad_norm": 0.8581610321998596,
449
+ "learning_rate": 2.9295608480314075e-05,
450
+ "loss": 0.7954,
451
+ "step": 315
452
+ },
453
+ {
454
+ "epoch": 0.4377564979480164,
455
+ "grad_norm": 0.9587123990058899,
456
+ "learning_rate": 2.9260946152211826e-05,
457
+ "loss": 0.7416,
458
+ "step": 320
459
+ },
460
+ {
461
+ "epoch": 0.44459644322845415,
462
+ "grad_norm": 1.002186894416809,
463
+ "learning_rate": 2.9225472849054066e-05,
464
+ "loss": 0.7843,
465
+ "step": 325
466
+ },
467
+ {
468
+ "epoch": 0.45143638850889195,
469
+ "grad_norm": 0.9037190079689026,
470
+ "learning_rate": 2.918919058809572e-05,
471
+ "loss": 0.7941,
472
+ "step": 330
473
+ },
474
+ {
475
+ "epoch": 0.4582763337893297,
476
+ "grad_norm": 1.0096290111541748,
477
+ "learning_rate": 2.9152101432594625e-05,
478
+ "loss": 0.7452,
479
+ "step": 335
480
+ },
481
+ {
482
+ "epoch": 0.46511627906976744,
483
+ "grad_norm": 0.8953573107719421,
484
+ "learning_rate": 2.911420749169412e-05,
485
+ "loss": 0.7744,
486
+ "step": 340
487
+ },
488
+ {
489
+ "epoch": 0.4719562243502052,
490
+ "grad_norm": 0.9449247717857361,
491
+ "learning_rate": 2.9075510920303195e-05,
492
+ "loss": 0.7543,
493
+ "step": 345
494
+ },
495
+ {
496
+ "epoch": 0.478796169630643,
497
+ "grad_norm": 0.9241800308227539,
498
+ "learning_rate": 2.903601391897388e-05,
499
+ "loss": 0.6998,
500
+ "step": 350
501
+ },
502
+ {
503
+ "epoch": 0.4856361149110807,
504
+ "grad_norm": 1.0948309898376465,
505
+ "learning_rate": 2.8995718733776148e-05,
506
+ "loss": 0.6661,
507
+ "step": 355
508
+ },
509
+ {
510
+ "epoch": 0.49247606019151846,
511
+ "grad_norm": 1.0500880479812622,
512
+ "learning_rate": 2.8954627656170162e-05,
513
+ "loss": 0.7215,
514
+ "step": 360
515
+ },
516
+ {
517
+ "epoch": 0.4993160054719562,
518
+ "grad_norm": 1.029900312423706,
519
+ "learning_rate": 2.891274302287598e-05,
520
+ "loss": 0.7165,
521
+ "step": 365
522
+ },
523
+ {
524
+ "epoch": 0.506155950752394,
525
+ "grad_norm": 0.9458325505256653,
526
+ "learning_rate": 2.8870067215740677e-05,
527
+ "loss": 0.6749,
528
+ "step": 370
529
+ },
530
+ {
531
+ "epoch": 0.5129958960328317,
532
+ "grad_norm": 0.9350931644439697,
533
+ "learning_rate": 2.882660266160288e-05,
534
+ "loss": 0.7208,
535
+ "step": 375
536
+ },
537
+ {
538
+ "epoch": 0.5198358413132695,
539
+ "grad_norm": 1.0697951316833496,
540
+ "learning_rate": 2.8782351832154786e-05,
541
+ "loss": 0.6869,
542
+ "step": 380
543
+ },
544
+ {
545
+ "epoch": 0.5266757865937073,
546
+ "grad_norm": 1.015529990196228,
547
+ "learning_rate": 2.873731724380157e-05,
548
+ "loss": 0.6879,
549
+ "step": 385
550
+ },
551
+ {
552
+ "epoch": 0.533515731874145,
553
+ "grad_norm": 0.9687970280647278,
554
+ "learning_rate": 2.869150145751833e-05,
555
+ "loss": 0.6883,
556
+ "step": 390
557
+ },
558
+ {
559
+ "epoch": 0.5403556771545828,
560
+ "grad_norm": 0.9908791184425354,
561
+ "learning_rate": 2.8644907078704404e-05,
562
+ "loss": 0.664,
563
+ "step": 395
564
+ },
565
+ {
566
+ "epoch": 0.5471956224350205,
567
+ "grad_norm": 0.9771539568901062,
568
+ "learning_rate": 2.8597536757035253e-05,
569
+ "loss": 0.673,
570
+ "step": 400
571
+ },
572
+ {
573
+ "epoch": 0.5540355677154583,
574
+ "grad_norm": 1.0252209901809692,
575
+ "learning_rate": 2.854939318631175e-05,
576
+ "loss": 0.6674,
577
+ "step": 405
578
+ },
579
+ {
580
+ "epoch": 0.560875512995896,
581
+ "grad_norm": 1.060739517211914,
582
+ "learning_rate": 2.8500479104307007e-05,
583
+ "loss": 0.6221,
584
+ "step": 410
585
+ },
586
+ {
587
+ "epoch": 0.5677154582763337,
588
+ "grad_norm": 1.0768077373504639,
589
+ "learning_rate": 2.8450797292610683e-05,
590
+ "loss": 0.6484,
591
+ "step": 415
592
+ },
593
+ {
594
+ "epoch": 0.5745554035567716,
595
+ "grad_norm": 1.0182243585586548,
596
+ "learning_rate": 2.840035057647081e-05,
597
+ "loss": 0.6087,
598
+ "step": 420
599
+ },
600
+ {
601
+ "epoch": 0.5813953488372093,
602
+ "grad_norm": 0.9320763349533081,
603
+ "learning_rate": 2.8349141824633122e-05,
604
+ "loss": 0.6438,
605
+ "step": 425
606
+ },
607
+ {
608
+ "epoch": 0.5882352941176471,
609
+ "grad_norm": 1.181852102279663,
610
+ "learning_rate": 2.8297173949177914e-05,
611
+ "loss": 0.6215,
612
+ "step": 430
613
+ },
614
+ {
615
+ "epoch": 0.5950752393980848,
616
+ "grad_norm": 1.0764853954315186,
617
+ "learning_rate": 2.824444990535446e-05,
618
+ "loss": 0.58,
619
+ "step": 435
620
+ },
621
+ {
622
+ "epoch": 0.6019151846785226,
623
+ "grad_norm": 0.9911195635795593,
624
+ "learning_rate": 2.8190972691412942e-05,
625
+ "loss": 0.6116,
626
+ "step": 440
627
+ },
628
+ {
629
+ "epoch": 0.6087551299589603,
630
+ "grad_norm": 0.9758164882659912,
631
+ "learning_rate": 2.813674534843394e-05,
632
+ "loss": 0.5763,
633
+ "step": 445
634
+ },
635
+ {
636
+ "epoch": 0.615595075239398,
637
+ "grad_norm": 1.1901512145996094,
638
+ "learning_rate": 2.808177096015553e-05,
639
+ "loss": 0.573,
640
+ "step": 450
641
+ },
642
+ {
643
+ "epoch": 0.6224350205198358,
644
+ "grad_norm": 1.116800308227539,
645
+ "learning_rate": 2.802605265279788e-05,
646
+ "loss": 0.5748,
647
+ "step": 455
648
+ },
649
+ {
650
+ "epoch": 0.6292749658002736,
651
+ "grad_norm": 1.0181605815887451,
652
+ "learning_rate": 2.7969593594885504e-05,
653
+ "loss": 0.5465,
654
+ "step": 460
655
+ },
656
+ {
657
+ "epoch": 0.6361149110807114,
658
+ "grad_norm": 1.2274715900421143,
659
+ "learning_rate": 2.7912396997067053e-05,
660
+ "loss": 0.5642,
661
+ "step": 465
662
+ },
663
+ {
664
+ "epoch": 0.6429548563611491,
665
+ "grad_norm": 1.013073205947876,
666
+ "learning_rate": 2.785446611193276e-05,
667
+ "loss": 0.5639,
668
+ "step": 470
669
+ },
670
+ {
671
+ "epoch": 0.6497948016415869,
672
+ "grad_norm": 1.0413159132003784,
673
+ "learning_rate": 2.7795804233829456e-05,
674
+ "loss": 0.563,
675
+ "step": 475
676
+ },
677
+ {
678
+ "epoch": 0.6566347469220246,
679
+ "grad_norm": 1.0619947910308838,
680
+ "learning_rate": 2.7736414698673245e-05,
681
+ "loss": 0.5703,
682
+ "step": 480
683
+ },
684
+ {
685
+ "epoch": 0.6634746922024624,
686
+ "grad_norm": 1.08983314037323,
687
+ "learning_rate": 2.7676300883759782e-05,
688
+ "loss": 0.5492,
689
+ "step": 485
690
+ },
691
+ {
692
+ "epoch": 0.6703146374829001,
693
+ "grad_norm": 1.145408272743225,
694
+ "learning_rate": 2.7615466207572238e-05,
695
+ "loss": 0.5372,
696
+ "step": 490
697
+ },
698
+ {
699
+ "epoch": 0.6771545827633378,
700
+ "grad_norm": 1.0494654178619385,
701
+ "learning_rate": 2.7553914129586885e-05,
702
+ "loss": 0.5505,
703
+ "step": 495
704
+ },
705
+ {
706
+ "epoch": 0.6839945280437757,
707
+ "grad_norm": 1.072014331817627,
708
+ "learning_rate": 2.7491648150076387e-05,
709
+ "loss": 0.5137,
710
+ "step": 500
711
+ },
712
+ {
713
+ "epoch": 0.6908344733242134,
714
+ "grad_norm": 1.064727783203125,
715
+ "learning_rate": 2.742867180991072e-05,
716
+ "loss": 0.5823,
717
+ "step": 505
718
+ },
719
+ {
720
+ "epoch": 0.6976744186046512,
721
+ "grad_norm": 1.164260745048523,
722
+ "learning_rate": 2.736498869035585e-05,
723
+ "loss": 0.5692,
724
+ "step": 510
725
+ },
726
+ {
727
+ "epoch": 0.7045143638850889,
728
+ "grad_norm": 1.066939353942871,
729
+ "learning_rate": 2.730060241287004e-05,
730
+ "loss": 0.5277,
731
+ "step": 515
732
+ },
733
+ {
734
+ "epoch": 0.7113543091655267,
735
+ "grad_norm": 1.0859402418136597,
736
+ "learning_rate": 2.723551663889795e-05,
737
+ "loss": 0.5122,
738
+ "step": 520
739
+ },
740
+ {
741
+ "epoch": 0.7181942544459644,
742
+ "grad_norm": 1.0558462142944336,
743
+ "learning_rate": 2.7169735069662372e-05,
744
+ "loss": 0.5227,
745
+ "step": 525
746
+ },
747
+ {
748
+ "epoch": 0.7250341997264022,
749
+ "grad_norm": 1.1296465396881104,
750
+ "learning_rate": 2.7103261445953812e-05,
751
+ "loss": 0.5292,
752
+ "step": 530
753
+ },
754
+ {
755
+ "epoch": 0.7318741450068399,
756
+ "grad_norm": 1.2444343566894531,
757
+ "learning_rate": 2.703609954791771e-05,
758
+ "loss": 0.5147,
759
+ "step": 535
760
+ },
761
+ {
762
+ "epoch": 0.7387140902872777,
763
+ "grad_norm": 1.1435099840164185,
764
+ "learning_rate": 2.69682531948395e-05,
765
+ "loss": 0.4856,
766
+ "step": 540
767
+ },
768
+ {
769
+ "epoch": 0.7455540355677155,
770
+ "grad_norm": 1.0850509405136108,
771
+ "learning_rate": 2.6899726244927438e-05,
772
+ "loss": 0.5221,
773
+ "step": 545
774
+ },
775
+ {
776
+ "epoch": 0.7523939808481532,
777
+ "grad_norm": 1.1499311923980713,
778
+ "learning_rate": 2.6830522595093145e-05,
779
+ "loss": 0.4908,
780
+ "step": 550
781
+ },
782
+ {
783
+ "epoch": 0.759233926128591,
784
+ "grad_norm": 1.2296826839447021,
785
+ "learning_rate": 2.6760646180730056e-05,
786
+ "loss": 0.5239,
787
+ "step": 555
788
+ },
789
+ {
790
+ "epoch": 0.7660738714090287,
791
+ "grad_norm": 1.0380018949508667,
792
+ "learning_rate": 2.6690100975489594e-05,
793
+ "loss": 0.5022,
794
+ "step": 560
795
+ },
796
+ {
797
+ "epoch": 0.7729138166894665,
798
+ "grad_norm": 1.0536307096481323,
799
+ "learning_rate": 2.6618890991055223e-05,
800
+ "loss": 0.482,
801
+ "step": 565
802
+ },
803
+ {
804
+ "epoch": 0.7797537619699042,
805
+ "grad_norm": 1.2291308641433716,
806
+ "learning_rate": 2.654702027691428e-05,
807
+ "loss": 0.4735,
808
+ "step": 570
809
+ },
810
+ {
811
+ "epoch": 0.786593707250342,
812
+ "grad_norm": 1.2991678714752197,
813
+ "learning_rate": 2.647449292012775e-05,
814
+ "loss": 0.4496,
815
+ "step": 575
816
+ },
817
+ {
818
+ "epoch": 0.7934336525307798,
819
+ "grad_norm": 1.1882619857788086,
820
+ "learning_rate": 2.6401313045097792e-05,
821
+ "loss": 0.4772,
822
+ "step": 580
823
+ },
824
+ {
825
+ "epoch": 0.8002735978112175,
826
+ "grad_norm": 1.3000277280807495,
827
+ "learning_rate": 2.632748481333322e-05,
828
+ "loss": 0.4492,
829
+ "step": 585
830
+ },
831
+ {
832
+ "epoch": 0.8071135430916553,
833
+ "grad_norm": 1.0775530338287354,
834
+ "learning_rate": 2.6253012423212874e-05,
835
+ "loss": 0.4448,
836
+ "step": 590
837
+ },
838
+ {
839
+ "epoch": 0.813953488372093,
840
+ "grad_norm": 1.0790055990219116,
841
+ "learning_rate": 2.617790010974682e-05,
842
+ "loss": 0.4467,
843
+ "step": 595
844
+ },
845
+ {
846
+ "epoch": 0.8207934336525308,
847
+ "grad_norm": 1.0407321453094482,
848
+ "learning_rate": 2.610215214433557e-05,
849
+ "loss": 0.4304,
850
+ "step": 600
851
+ },
852
+ {
853
+ "epoch": 0.8276333789329685,
854
+ "grad_norm": 1.2868131399154663,
855
+ "learning_rate": 2.602577283452715e-05,
856
+ "loss": 0.5008,
857
+ "step": 605
858
+ },
859
+ {
860
+ "epoch": 0.8344733242134063,
861
+ "grad_norm": 1.0994542837142944,
862
+ "learning_rate": 2.594876652377215e-05,
863
+ "loss": 0.4496,
864
+ "step": 610
865
+ },
866
+ {
867
+ "epoch": 0.841313269493844,
868
+ "grad_norm": 1.1733685731887817,
869
+ "learning_rate": 2.5871137591176733e-05,
870
+ "loss": 0.4544,
871
+ "step": 615
872
+ },
873
+ {
874
+ "epoch": 0.8481532147742818,
875
+ "grad_norm": 1.1335357427597046,
876
+ "learning_rate": 2.57928904512536e-05,
877
+ "loss": 0.4317,
878
+ "step": 620
879
+ },
880
+ {
881
+ "epoch": 0.8549931600547196,
882
+ "grad_norm": 1.037997841835022,
883
+ "learning_rate": 2.571402955367095e-05,
884
+ "loss": 0.4379,
885
+ "step": 625
886
+ },
887
+ {
888
+ "epoch": 0.8618331053351573,
889
+ "grad_norm": 1.2495006322860718,
890
+ "learning_rate": 2.563455938299945e-05,
891
+ "loss": 0.4417,
892
+ "step": 630
893
+ },
894
+ {
895
+ "epoch": 0.8686730506155951,
896
+ "grad_norm": 1.200516939163208,
897
+ "learning_rate": 2.5554484458457213e-05,
898
+ "loss": 0.3971,
899
+ "step": 635
900
+ },
901
+ {
902
+ "epoch": 0.8755129958960328,
903
+ "grad_norm": 1.1400073766708374,
904
+ "learning_rate": 2.5473809333652788e-05,
905
+ "loss": 0.3895,
906
+ "step": 640
907
+ },
908
+ {
909
+ "epoch": 0.8823529411764706,
910
+ "grad_norm": 1.005179762840271,
911
+ "learning_rate": 2.5392538596326226e-05,
912
+ "loss": 0.4198,
913
+ "step": 645
914
+ },
915
+ {
916
+ "epoch": 0.8891928864569083,
917
+ "grad_norm": 1.2099529504776,
918
+ "learning_rate": 2.5310676868088183e-05,
919
+ "loss": 0.4214,
920
+ "step": 650
921
+ },
922
+ {
923
+ "epoch": 0.896032831737346,
924
+ "grad_norm": 1.0562937259674072,
925
+ "learning_rate": 2.5228228804157105e-05,
926
+ "loss": 0.3988,
927
+ "step": 655
928
+ },
929
+ {
930
+ "epoch": 0.9028727770177839,
931
+ "grad_norm": 1.0145800113677979,
932
+ "learning_rate": 2.5145199093094506e-05,
933
+ "loss": 0.4154,
934
+ "step": 660
935
+ },
936
+ {
937
+ "epoch": 0.9097127222982216,
938
+ "grad_norm": 1.2560648918151855,
939
+ "learning_rate": 2.5061592456538333e-05,
940
+ "loss": 0.4108,
941
+ "step": 665
942
+ },
943
+ {
944
+ "epoch": 0.9165526675786594,
945
+ "grad_norm": 1.1360726356506348,
946
+ "learning_rate": 2.4977413648934467e-05,
947
+ "loss": 0.4046,
948
+ "step": 670
949
+ },
950
+ {
951
+ "epoch": 0.9233926128590971,
952
+ "grad_norm": 1.268717646598816,
953
+ "learning_rate": 2.4892667457266362e-05,
954
+ "loss": 0.4413,
955
+ "step": 675
956
+ },
957
+ {
958
+ "epoch": 0.9302325581395349,
959
+ "grad_norm": 1.0952569246292114,
960
+ "learning_rate": 2.4807358700782796e-05,
961
+ "loss": 0.3929,
962
+ "step": 680
963
+ },
964
+ {
965
+ "epoch": 0.9370725034199726,
966
+ "grad_norm": 1.1098562479019165,
967
+ "learning_rate": 2.4721492230723855e-05,
968
+ "loss": 0.3729,
969
+ "step": 685
970
+ },
971
+ {
972
+ "epoch": 0.9439124487004104,
973
+ "grad_norm": 1.133565902709961,
974
+ "learning_rate": 2.4635072930045033e-05,
975
+ "loss": 0.4124,
976
+ "step": 690
977
+ },
978
+ {
979
+ "epoch": 0.9507523939808481,
980
+ "grad_norm": 1.1997113227844238,
981
+ "learning_rate": 2.454810571313955e-05,
982
+ "loss": 0.3792,
983
+ "step": 695
984
+ },
985
+ {
986
+ "epoch": 0.957592339261286,
987
+ "grad_norm": 1.1516748666763306,
988
+ "learning_rate": 2.4460595525558903e-05,
989
+ "loss": 0.3954,
990
+ "step": 700
991
+ },
992
+ {
993
+ "epoch": 0.9644322845417237,
994
+ "grad_norm": 1.3405410051345825,
995
+ "learning_rate": 2.437254734373162e-05,
996
+ "loss": 0.3484,
997
+ "step": 705
998
+ },
999
+ {
1000
+ "epoch": 0.9712722298221614,
1001
+ "grad_norm": 1.0757001638412476,
1002
+ "learning_rate": 2.4283966174680254e-05,
1003
+ "loss": 0.388,
1004
+ "step": 710
1005
+ },
1006
+ {
1007
+ "epoch": 0.9781121751025992,
1008
+ "grad_norm": 0.99599289894104,
1009
+ "learning_rate": 2.4194857055736682e-05,
1010
+ "loss": 0.382,
1011
+ "step": 715
1012
+ },
1013
+ {
1014
+ "epoch": 0.9849521203830369,
1015
+ "grad_norm": 1.2733392715454102,
1016
+ "learning_rate": 2.4105225054255607e-05,
1017
+ "loss": 0.3769,
1018
+ "step": 720
1019
+ },
1020
+ {
1021
+ "epoch": 0.9917920656634747,
1022
+ "grad_norm": 1.2212460041046143,
1023
+ "learning_rate": 2.4015075267326432e-05,
1024
+ "loss": 0.3501,
1025
+ "step": 725
1026
+ },
1027
+ {
1028
+ "epoch": 0.9986320109439124,
1029
+ "grad_norm": 1.194299578666687,
1030
+ "learning_rate": 2.392441282148337e-05,
1031
+ "loss": 0.3182,
1032
+ "step": 730
1033
+ },
1034
+ {
1035
+ "epoch": 1.0054719562243501,
1036
+ "grad_norm": 1.142229437828064,
1037
+ "learning_rate": 2.383324287241394e-05,
1038
+ "loss": 0.3078,
1039
+ "step": 735
1040
+ },
1041
+ {
1042
+ "epoch": 1.012311901504788,
1043
+ "grad_norm": 1.1372795104980469,
1044
+ "learning_rate": 2.3741570604665767e-05,
1045
+ "loss": 0.2853,
1046
+ "step": 740
1047
+ },
1048
+ {
1049
+ "epoch": 1.0191518467852256,
1050
+ "grad_norm": 1.0447818040847778,
1051
+ "learning_rate": 2.3649401231351752e-05,
1052
+ "loss": 0.2957,
1053
+ "step": 745
1054
+ },
1055
+ {
1056
+ "epoch": 1.0259917920656634,
1057
+ "grad_norm": 1.1684584617614746,
1058
+ "learning_rate": 2.3556739993853635e-05,
1059
+ "loss": 0.3014,
1060
+ "step": 750
1061
+ },
1062
+ {
1063
+ "epoch": 1.0328317373461013,
1064
+ "grad_norm": 1.0782568454742432,
1065
+ "learning_rate": 2.346359216152391e-05,
1066
+ "loss": 0.273,
1067
+ "step": 755
1068
+ },
1069
+ {
1070
+ "epoch": 1.039671682626539,
1071
+ "grad_norm": 1.015937089920044,
1072
+ "learning_rate": 2.3369963031386186e-05,
1073
+ "loss": 0.2725,
1074
+ "step": 760
1075
+ },
1076
+ {
1077
+ "epoch": 1.0465116279069768,
1078
+ "grad_norm": 0.9925934672355652,
1079
+ "learning_rate": 2.3275857927833962e-05,
1080
+ "loss": 0.2782,
1081
+ "step": 765
1082
+ },
1083
+ {
1084
+ "epoch": 1.0533515731874146,
1085
+ "grad_norm": 1.1000502109527588,
1086
+ "learning_rate": 2.318128220232786e-05,
1087
+ "loss": 0.2868,
1088
+ "step": 770
1089
+ },
1090
+ {
1091
+ "epoch": 1.0601915184678523,
1092
+ "grad_norm": 1.124937653541565,
1093
+ "learning_rate": 2.3086241233091258e-05,
1094
+ "loss": 0.2844,
1095
+ "step": 775
1096
+ },
1097
+ {
1098
+ "epoch": 1.06703146374829,
1099
+ "grad_norm": 1.2203454971313477,
1100
+ "learning_rate": 2.299074042480451e-05,
1101
+ "loss": 0.2798,
1102
+ "step": 780
1103
+ },
1104
+ {
1105
+ "epoch": 1.0738714090287278,
1106
+ "grad_norm": 1.0879415273666382,
1107
+ "learning_rate": 2.2894785208297547e-05,
1108
+ "loss": 0.2586,
1109
+ "step": 785
1110
+ },
1111
+ {
1112
+ "epoch": 1.0807113543091655,
1113
+ "grad_norm": 1.1645656824111938,
1114
+ "learning_rate": 2.279838104024107e-05,
1115
+ "loss": 0.277,
1116
+ "step": 790
1117
+ },
1118
+ {
1119
+ "epoch": 1.0875512995896033,
1120
+ "grad_norm": 1.0477960109710693,
1121
+ "learning_rate": 2.270153340283624e-05,
1122
+ "loss": 0.291,
1123
+ "step": 795
1124
+ },
1125
+ {
1126
+ "epoch": 1.094391244870041,
1127
+ "grad_norm": 1.1401909589767456,
1128
+ "learning_rate": 2.2604247803502926e-05,
1129
+ "loss": 0.2913,
1130
+ "step": 800
1131
+ },
1132
+ {
1133
+ "epoch": 1.1012311901504788,
1134
+ "grad_norm": 1.1064962148666382,
1135
+ "learning_rate": 2.2506529774566495e-05,
1136
+ "loss": 0.2393,
1137
+ "step": 805
1138
+ },
1139
+ {
1140
+ "epoch": 1.1080711354309165,
1141
+ "grad_norm": 1.2653497457504272,
1142
+ "learning_rate": 2.2408384872943245e-05,
1143
+ "loss": 0.2544,
1144
+ "step": 810
1145
+ },
1146
+ {
1147
+ "epoch": 1.1149110807113543,
1148
+ "grad_norm": 1.1752402782440186,
1149
+ "learning_rate": 2.2309818679824365e-05,
1150
+ "loss": 0.2607,
1151
+ "step": 815
1152
+ },
1153
+ {
1154
+ "epoch": 1.121751025991792,
1155
+ "grad_norm": 1.1659537553787231,
1156
+ "learning_rate": 2.221083680035856e-05,
1157
+ "loss": 0.2904,
1158
+ "step": 820
1159
+ },
1160
+ {
1161
+ "epoch": 1.1285909712722297,
1162
+ "grad_norm": 1.0146831274032593,
1163
+ "learning_rate": 2.2111444863333317e-05,
1164
+ "loss": 0.2847,
1165
+ "step": 825
1166
+ },
1167
+ {
1168
+ "epoch": 1.1354309165526675,
1169
+ "grad_norm": 1.1357191801071167,
1170
+ "learning_rate": 2.2011648520854805e-05,
1171
+ "loss": 0.2684,
1172
+ "step": 830
1173
+ },
1174
+ {
1175
+ "epoch": 1.1422708618331052,
1176
+ "grad_norm": 1.0175282955169678,
1177
+ "learning_rate": 2.191145344802644e-05,
1178
+ "loss": 0.2871,
1179
+ "step": 835
1180
+ },
1181
+ {
1182
+ "epoch": 1.1491108071135432,
1183
+ "grad_norm": 0.9281819462776184,
1184
+ "learning_rate": 2.181086534262619e-05,
1185
+ "loss": 0.2551,
1186
+ "step": 840
1187
+ },
1188
+ {
1189
+ "epoch": 1.155950752393981,
1190
+ "grad_norm": 1.1165109872817993,
1191
+ "learning_rate": 2.170988992478255e-05,
1192
+ "loss": 0.2583,
1193
+ "step": 845
1194
+ },
1195
+ {
1196
+ "epoch": 1.1627906976744187,
1197
+ "grad_norm": 1.1433641910552979,
1198
+ "learning_rate": 2.1608532936649242e-05,
1199
+ "loss": 0.2594,
1200
+ "step": 850
1201
+ },
1202
+ {
1203
+ "epoch": 1.1696306429548564,
1204
+ "grad_norm": 1.181136965751648,
1205
+ "learning_rate": 2.1506800142078693e-05,
1206
+ "loss": 0.2769,
1207
+ "step": 855
1208
+ },
1209
+ {
1210
+ "epoch": 1.1764705882352942,
1211
+ "grad_norm": 1.179027795791626,
1212
+ "learning_rate": 2.140469732629426e-05,
1213
+ "loss": 0.2554,
1214
+ "step": 860
1215
+ },
1216
+ {
1217
+ "epoch": 1.183310533515732,
1218
+ "grad_norm": 1.1550763845443726,
1219
+ "learning_rate": 2.1302230295561226e-05,
1220
+ "loss": 0.2227,
1221
+ "step": 865
1222
+ },
1223
+ {
1224
+ "epoch": 1.1901504787961696,
1225
+ "grad_norm": 1.1192501783370972,
1226
+ "learning_rate": 2.119940487685665e-05,
1227
+ "loss": 0.2423,
1228
+ "step": 870
1229
+ },
1230
+ {
1231
+ "epoch": 1.1969904240766074,
1232
+ "grad_norm": 0.970893144607544,
1233
+ "learning_rate": 2.1096226917537975e-05,
1234
+ "loss": 0.2596,
1235
+ "step": 875
1236
+ },
1237
+ {
1238
+ "epoch": 1.2038303693570451,
1239
+ "grad_norm": 1.10446298122406,
1240
+ "learning_rate": 2.0992702285010515e-05,
1241
+ "loss": 0.2484,
1242
+ "step": 880
1243
+ },
1244
+ {
1245
+ "epoch": 1.2106703146374829,
1246
+ "grad_norm": 1.209159016609192,
1247
+ "learning_rate": 2.0888836866393807e-05,
1248
+ "loss": 0.238,
1249
+ "step": 885
1250
+ },
1251
+ {
1252
+ "epoch": 1.2175102599179206,
1253
+ "grad_norm": 1.2407492399215698,
1254
+ "learning_rate": 2.0784636568186807e-05,
1255
+ "loss": 0.2405,
1256
+ "step": 890
1257
+ },
1258
+ {
1259
+ "epoch": 1.2243502051983584,
1260
+ "grad_norm": 1.2051174640655518,
1261
+ "learning_rate": 2.0680107315932018e-05,
1262
+ "loss": 0.2484,
1263
+ "step": 895
1264
+ },
1265
+ {
1266
+ "epoch": 1.231190150478796,
1267
+ "grad_norm": 1.0806186199188232,
1268
+ "learning_rate": 2.0575255053878532e-05,
1269
+ "loss": 0.2213,
1270
+ "step": 900
1271
+ },
1272
+ {
1273
+ "epoch": 1.2380300957592338,
1274
+ "grad_norm": 1.2482553720474243,
1275
+ "learning_rate": 2.0470085744643992e-05,
1276
+ "loss": 0.234,
1277
+ "step": 905
1278
+ },
1279
+ {
1280
+ "epoch": 1.2448700410396718,
1281
+ "grad_norm": 1.0524805784225464,
1282
+ "learning_rate": 2.0364605368875498e-05,
1283
+ "loss": 0.2368,
1284
+ "step": 910
1285
+ },
1286
+ {
1287
+ "epoch": 1.2517099863201095,
1288
+ "grad_norm": 1.1557000875473022,
1289
+ "learning_rate": 2.0258819924909538e-05,
1290
+ "loss": 0.2379,
1291
+ "step": 915
1292
+ },
1293
+ {
1294
+ "epoch": 1.2585499316005473,
1295
+ "grad_norm": 1.1021133661270142,
1296
+ "learning_rate": 2.0152735428430864e-05,
1297
+ "loss": 0.2204,
1298
+ "step": 920
1299
+ },
1300
+ {
1301
+ "epoch": 1.265389876880985,
1302
+ "grad_norm": 1.1099745035171509,
1303
+ "learning_rate": 2.004635791213039e-05,
1304
+ "loss": 0.2363,
1305
+ "step": 925
1306
+ },
1307
+ {
1308
+ "epoch": 1.2722298221614228,
1309
+ "grad_norm": 0.9470094442367554,
1310
+ "learning_rate": 1.9939693425362168e-05,
1311
+ "loss": 0.233,
1312
+ "step": 930
1313
+ },
1314
+ {
1315
+ "epoch": 1.2790697674418605,
1316
+ "grad_norm": 1.0972777605056763,
1317
+ "learning_rate": 1.9832748033799334e-05,
1318
+ "loss": 0.2397,
1319
+ "step": 935
1320
+ },
1321
+ {
1322
+ "epoch": 1.2859097127222983,
1323
+ "grad_norm": 1.125274658203125,
1324
+ "learning_rate": 1.97255278190892e-05,
1325
+ "loss": 0.2359,
1326
+ "step": 940
1327
+ },
1328
+ {
1329
+ "epoch": 1.292749658002736,
1330
+ "grad_norm": 1.126130223274231,
1331
+ "learning_rate": 1.9618038878507412e-05,
1332
+ "loss": 0.2201,
1333
+ "step": 945
1334
+ },
1335
+ {
1336
+ "epoch": 1.2995896032831737,
1337
+ "grad_norm": 1.1277283430099487,
1338
+ "learning_rate": 1.95102873246112e-05,
1339
+ "loss": 0.2136,
1340
+ "step": 950
1341
+ },
1342
+ {
1343
+ "epoch": 1.3064295485636115,
1344
+ "grad_norm": 0.9474350214004517,
1345
+ "learning_rate": 1.9402279284891785e-05,
1346
+ "loss": 0.1938,
1347
+ "step": 955
1348
+ },
1349
+ {
1350
+ "epoch": 1.3132694938440492,
1351
+ "grad_norm": 0.8682848811149597,
1352
+ "learning_rate": 1.929402090142594e-05,
1353
+ "loss": 0.2363,
1354
+ "step": 960
1355
+ },
1356
+ {
1357
+ "epoch": 1.320109439124487,
1358
+ "grad_norm": 1.2367846965789795,
1359
+ "learning_rate": 1.9185518330526687e-05,
1360
+ "loss": 0.2296,
1361
+ "step": 965
1362
+ },
1363
+ {
1364
+ "epoch": 1.3269493844049247,
1365
+ "grad_norm": 1.2060308456420898,
1366
+ "learning_rate": 1.907677774239321e-05,
1367
+ "loss": 0.1997,
1368
+ "step": 970
1369
+ },
1370
+ {
1371
+ "epoch": 1.3337893296853625,
1372
+ "grad_norm": 1.0779154300689697,
1373
+ "learning_rate": 1.896780532076001e-05,
1374
+ "loss": 0.2114,
1375
+ "step": 975
1376
+ },
1377
+ {
1378
+ "epoch": 1.3406292749658002,
1379
+ "grad_norm": 1.1252628564834595,
1380
+ "learning_rate": 1.88586072625452e-05,
1381
+ "loss": 0.2134,
1382
+ "step": 980
1383
+ },
1384
+ {
1385
+ "epoch": 1.347469220246238,
1386
+ "grad_norm": 1.1493712663650513,
1387
+ "learning_rate": 1.8749189777498146e-05,
1388
+ "loss": 0.2211,
1389
+ "step": 985
1390
+ },
1391
+ {
1392
+ "epoch": 1.3543091655266757,
1393
+ "grad_norm": 1.0766429901123047,
1394
+ "learning_rate": 1.863955908784633e-05,
1395
+ "loss": 0.1893,
1396
+ "step": 990
1397
+ },
1398
+ {
1399
+ "epoch": 1.3611491108071134,
1400
+ "grad_norm": 1.1063032150268555,
1401
+ "learning_rate": 1.8529721427941506e-05,
1402
+ "loss": 0.2368,
1403
+ "step": 995
1404
+ },
1405
+ {
1406
+ "epoch": 1.3679890560875512,
1407
+ "grad_norm": 1.0775789022445679,
1408
+ "learning_rate": 1.8419683043905166e-05,
1409
+ "loss": 0.2186,
1410
+ "step": 1000
1411
+ },
1412
+ {
1413
+ "epoch": 1.3748290013679891,
1414
+ "grad_norm": 1.2360905408859253,
1415
+ "learning_rate": 1.830945019327336e-05,
1416
+ "loss": 0.206,
1417
+ "step": 1005
1418
+ },
1419
+ {
1420
+ "epoch": 1.3816689466484269,
1421
+ "grad_norm": 1.123159646987915,
1422
+ "learning_rate": 1.8199029144640833e-05,
1423
+ "loss": 0.1942,
1424
+ "step": 1010
1425
+ },
1426
+ {
1427
+ "epoch": 1.3885088919288646,
1428
+ "grad_norm": 1.119447946548462,
1429
+ "learning_rate": 1.8088426177304564e-05,
1430
+ "loss": 0.1901,
1431
+ "step": 1015
1432
+ },
1433
+ {
1434
+ "epoch": 1.3953488372093024,
1435
+ "grad_norm": 1.341801404953003,
1436
+ "learning_rate": 1.797764758090667e-05,
1437
+ "loss": 0.2024,
1438
+ "step": 1020
1439
+ },
1440
+ {
1441
+ "epoch": 1.40218878248974,
1442
+ "grad_norm": 1.0289816856384277,
1443
+ "learning_rate": 1.786669965507674e-05,
1444
+ "loss": 0.1737,
1445
+ "step": 1025
1446
+ },
1447
+ {
1448
+ "epoch": 1.4090287277701778,
1449
+ "grad_norm": 1.4920783042907715,
1450
+ "learning_rate": 1.7755588709073605e-05,
1451
+ "loss": 0.169,
1452
+ "step": 1030
1453
+ },
1454
+ {
1455
+ "epoch": 1.4158686730506156,
1456
+ "grad_norm": 1.148211121559143,
1457
+ "learning_rate": 1.7644321061426518e-05,
1458
+ "loss": 0.1983,
1459
+ "step": 1035
1460
+ },
1461
+ {
1462
+ "epoch": 1.4227086183310533,
1463
+ "grad_norm": 1.2545026540756226,
1464
+ "learning_rate": 1.753290303957587e-05,
1465
+ "loss": 0.2245,
1466
+ "step": 1040
1467
+ },
1468
+ {
1469
+ "epoch": 1.429548563611491,
1470
+ "grad_norm": 0.9977272748947144,
1471
+ "learning_rate": 1.7421340979513374e-05,
1472
+ "loss": 0.2006,
1473
+ "step": 1045
1474
+ },
1475
+ {
1476
+ "epoch": 1.4363885088919288,
1477
+ "grad_norm": 1.2933138608932495,
1478
+ "learning_rate": 1.730964122542172e-05,
1479
+ "loss": 0.1865,
1480
+ "step": 1050
1481
+ },
1482
+ {
1483
+ "epoch": 1.4432284541723666,
1484
+ "grad_norm": 1.2291208505630493,
1485
+ "learning_rate": 1.719781012931383e-05,
1486
+ "loss": 0.1827,
1487
+ "step": 1055
1488
+ },
1489
+ {
1490
+ "epoch": 1.4500683994528043,
1491
+ "grad_norm": 1.0998331308364868,
1492
+ "learning_rate": 1.7085854050671648e-05,
1493
+ "loss": 0.1895,
1494
+ "step": 1060
1495
+ },
1496
+ {
1497
+ "epoch": 1.4569083447332423,
1498
+ "grad_norm": 1.1165440082550049,
1499
+ "learning_rate": 1.6973779356084454e-05,
1500
+ "loss": 0.1904,
1501
+ "step": 1065
1502
+ },
1503
+ {
1504
+ "epoch": 1.46374829001368,
1505
+ "grad_norm": 1.2620481252670288,
1506
+ "learning_rate": 1.6861592418886862e-05,
1507
+ "loss": 0.1887,
1508
+ "step": 1070
1509
+ },
1510
+ {
1511
+ "epoch": 1.4705882352941178,
1512
+ "grad_norm": 1.2061619758605957,
1513
+ "learning_rate": 1.674929961879637e-05,
1514
+ "loss": 0.1799,
1515
+ "step": 1075
1516
+ },
1517
+ {
1518
+ "epoch": 1.4774281805745555,
1519
+ "grad_norm": 1.1428618431091309,
1520
+ "learning_rate": 1.6636907341550544e-05,
1521
+ "loss": 0.1655,
1522
+ "step": 1080
1523
+ },
1524
+ {
1525
+ "epoch": 1.4842681258549932,
1526
+ "grad_norm": 1.0828522443771362,
1527
+ "learning_rate": 1.652442197854393e-05,
1528
+ "loss": 0.1716,
1529
+ "step": 1085
1530
+ },
1531
+ {
1532
+ "epoch": 1.491108071135431,
1533
+ "grad_norm": 1.0004147291183472,
1534
+ "learning_rate": 1.6411849926464555e-05,
1535
+ "loss": 0.1922,
1536
+ "step": 1090
1537
+ },
1538
+ {
1539
+ "epoch": 1.4979480164158687,
1540
+ "grad_norm": 1.1220521926879883,
1541
+ "learning_rate": 1.6299197586930176e-05,
1542
+ "loss": 0.1974,
1543
+ "step": 1095
1544
+ },
1545
+ {
1546
+ "epoch": 1.5047879616963065,
1547
+ "grad_norm": 1.0169928073883057,
1548
+ "learning_rate": 1.6186471366124248e-05,
1549
+ "loss": 0.1929,
1550
+ "step": 1100
1551
+ },
1552
+ {
1553
+ "epoch": 1.5116279069767442,
1554
+ "grad_norm": 1.0538076162338257,
1555
+ "learning_rate": 1.607367767443163e-05,
1556
+ "loss": 0.1823,
1557
+ "step": 1105
1558
+ },
1559
+ {
1560
+ "epoch": 1.518467852257182,
1561
+ "grad_norm": 1.1773090362548828,
1562
+ "learning_rate": 1.5960822926074018e-05,
1563
+ "loss": 0.1956,
1564
+ "step": 1110
1565
+ },
1566
+ {
1567
+ "epoch": 1.5253077975376197,
1568
+ "grad_norm": 1.0526224374771118,
1569
+ "learning_rate": 1.5847913538745226e-05,
1570
+ "loss": 0.1685,
1571
+ "step": 1115
1572
+ },
1573
+ {
1574
+ "epoch": 1.5321477428180574,
1575
+ "grad_norm": 1.0106682777404785,
1576
+ "learning_rate": 1.573495593324622e-05,
1577
+ "loss": 0.1844,
1578
+ "step": 1120
1579
+ },
1580
+ {
1581
+ "epoch": 1.5389876880984952,
1582
+ "grad_norm": 1.4278212785720825,
1583
+ "learning_rate": 1.5621956533119942e-05,
1584
+ "loss": 0.1565,
1585
+ "step": 1125
1586
+ },
1587
+ {
1588
+ "epoch": 1.545827633378933,
1589
+ "grad_norm": 1.022054672241211,
1590
+ "learning_rate": 1.5508921764286113e-05,
1591
+ "loss": 0.1597,
1592
+ "step": 1130
1593
+ },
1594
+ {
1595
+ "epoch": 1.5526675786593707,
1596
+ "grad_norm": 0.9699917435646057,
1597
+ "learning_rate": 1.5395858054675727e-05,
1598
+ "loss": 0.1937,
1599
+ "step": 1135
1600
+ },
1601
+ {
1602
+ "epoch": 1.5595075239398084,
1603
+ "grad_norm": 1.0063931941986084,
1604
+ "learning_rate": 1.528277183386556e-05,
1605
+ "loss": 0.1784,
1606
+ "step": 1140
1607
+ },
1608
+ {
1609
+ "epoch": 1.5663474692202461,
1610
+ "grad_norm": 1.1803474426269531,
1611
+ "learning_rate": 1.5169669532712534e-05,
1612
+ "loss": 0.185,
1613
+ "step": 1145
1614
+ },
1615
+ {
1616
+ "epoch": 1.573187414500684,
1617
+ "grad_norm": 0.9540643692016602,
1618
+ "learning_rate": 1.5056557582988002e-05,
1619
+ "loss": 0.1657,
1620
+ "step": 1150
1621
+ },
1622
+ {
1623
+ "epoch": 1.5800273597811216,
1624
+ "grad_norm": 1.0490425825119019,
1625
+ "learning_rate": 1.4943442417011999e-05,
1626
+ "loss": 0.1426,
1627
+ "step": 1155
1628
+ },
1629
+ {
1630
+ "epoch": 1.5868673050615594,
1631
+ "grad_norm": 1.0945137739181519,
1632
+ "learning_rate": 1.4830330467287463e-05,
1633
+ "loss": 0.156,
1634
+ "step": 1160
1635
+ },
1636
+ {
1637
+ "epoch": 1.5937072503419971,
1638
+ "grad_norm": 1.0034267902374268,
1639
+ "learning_rate": 1.4717228166134439e-05,
1640
+ "loss": 0.1562,
1641
+ "step": 1165
1642
+ },
1643
+ {
1644
+ "epoch": 1.6005471956224349,
1645
+ "grad_norm": 0.9562016725540161,
1646
+ "learning_rate": 1.4604141945324274e-05,
1647
+ "loss": 0.1539,
1648
+ "step": 1170
1649
+ },
1650
+ {
1651
+ "epoch": 1.6073871409028728,
1652
+ "grad_norm": 1.1179293394088745,
1653
+ "learning_rate": 1.4491078235713891e-05,
1654
+ "loss": 0.1536,
1655
+ "step": 1175
1656
+ },
1657
+ {
1658
+ "epoch": 1.6142270861833106,
1659
+ "grad_norm": 0.9614067673683167,
1660
+ "learning_rate": 1.437804346688006e-05,
1661
+ "loss": 0.1537,
1662
+ "step": 1180
1663
+ },
1664
+ {
1665
+ "epoch": 1.6210670314637483,
1666
+ "grad_norm": 1.1113111972808838,
1667
+ "learning_rate": 1.4265044066753788e-05,
1668
+ "loss": 0.1651,
1669
+ "step": 1185
1670
+ },
1671
+ {
1672
+ "epoch": 1.627906976744186,
1673
+ "grad_norm": 0.9136995673179626,
1674
+ "learning_rate": 1.4152086461254773e-05,
1675
+ "loss": 0.1609,
1676
+ "step": 1190
1677
+ },
1678
+ {
1679
+ "epoch": 1.6347469220246238,
1680
+ "grad_norm": 1.0922030210494995,
1681
+ "learning_rate": 1.4039177073925983e-05,
1682
+ "loss": 0.1297,
1683
+ "step": 1195
1684
+ },
1685
+ {
1686
+ "epoch": 1.6415868673050615,
1687
+ "grad_norm": 1.0154660940170288,
1688
+ "learning_rate": 1.3926322325568375e-05,
1689
+ "loss": 0.1524,
1690
+ "step": 1200
1691
+ },
1692
+ {
1693
+ "epoch": 1.6484268125854993,
1694
+ "grad_norm": 0.843471348285675,
1695
+ "learning_rate": 1.3813528633875757e-05,
1696
+ "loss": 0.1313,
1697
+ "step": 1205
1698
+ },
1699
+ {
1700
+ "epoch": 1.655266757865937,
1701
+ "grad_norm": 0.9563076496124268,
1702
+ "learning_rate": 1.3700802413069828e-05,
1703
+ "loss": 0.1328,
1704
+ "step": 1210
1705
+ },
1706
+ {
1707
+ "epoch": 1.662106703146375,
1708
+ "grad_norm": 1.0425236225128174,
1709
+ "learning_rate": 1.3588150073535444e-05,
1710
+ "loss": 0.1331,
1711
+ "step": 1215
1712
+ },
1713
+ {
1714
+ "epoch": 1.6689466484268127,
1715
+ "grad_norm": 1.05912446975708,
1716
+ "learning_rate": 1.347557802145607e-05,
1717
+ "loss": 0.1499,
1718
+ "step": 1220
1719
+ },
1720
+ {
1721
+ "epoch": 1.6757865937072505,
1722
+ "grad_norm": 1.124664068222046,
1723
+ "learning_rate": 1.3363092658449458e-05,
1724
+ "loss": 0.1616,
1725
+ "step": 1225
1726
+ },
1727
+ {
1728
+ "epoch": 1.6826265389876882,
1729
+ "grad_norm": 0.9814010262489319,
1730
+ "learning_rate": 1.3250700381203634e-05,
1731
+ "loss": 0.1335,
1732
+ "step": 1230
1733
+ },
1734
+ {
1735
+ "epoch": 1.689466484268126,
1736
+ "grad_norm": 1.013527750968933,
1737
+ "learning_rate": 1.313840758111314e-05,
1738
+ "loss": 0.1209,
1739
+ "step": 1235
1740
+ },
1741
+ {
1742
+ "epoch": 1.6963064295485637,
1743
+ "grad_norm": 0.968645453453064,
1744
+ "learning_rate": 1.3026220643915548e-05,
1745
+ "loss": 0.1458,
1746
+ "step": 1240
1747
+ },
1748
+ {
1749
+ "epoch": 1.7031463748290014,
1750
+ "grad_norm": 1.0528912544250488,
1751
+ "learning_rate": 1.2914145949328356e-05,
1752
+ "loss": 0.1337,
1753
+ "step": 1245
1754
+ },
1755
+ {
1756
+ "epoch": 1.7099863201094392,
1757
+ "grad_norm": 1.0264033079147339,
1758
+ "learning_rate": 1.2802189870686172e-05,
1759
+ "loss": 0.1603,
1760
+ "step": 1250
1761
+ },
1762
+ {
1763
+ "epoch": 1.716826265389877,
1764
+ "grad_norm": 1.0151350498199463,
1765
+ "learning_rate": 1.2690358774578284e-05,
1766
+ "loss": 0.1323,
1767
+ "step": 1255
1768
+ },
1769
+ {
1770
+ "epoch": 1.7236662106703147,
1771
+ "grad_norm": 1.1355870962142944,
1772
+ "learning_rate": 1.257865902048663e-05,
1773
+ "loss": 0.1165,
1774
+ "step": 1260
1775
+ },
1776
+ {
1777
+ "epoch": 1.7305061559507524,
1778
+ "grad_norm": 0.8444592356681824,
1779
+ "learning_rate": 1.2467096960424134e-05,
1780
+ "loss": 0.1352,
1781
+ "step": 1265
1782
+ },
1783
+ {
1784
+ "epoch": 1.7373461012311902,
1785
+ "grad_norm": 0.8394283652305603,
1786
+ "learning_rate": 1.2355678938573491e-05,
1787
+ "loss": 0.1219,
1788
+ "step": 1270
1789
+ },
1790
+ {
1791
+ "epoch": 1.744186046511628,
1792
+ "grad_norm": 0.8611863255500793,
1793
+ "learning_rate": 1.2244411290926398e-05,
1794
+ "loss": 0.1301,
1795
+ "step": 1275
1796
+ },
1797
+ {
1798
+ "epoch": 1.7510259917920656,
1799
+ "grad_norm": 0.9093337059020996,
1800
+ "learning_rate": 1.2133300344923259e-05,
1801
+ "loss": 0.1338,
1802
+ "step": 1280
1803
+ },
1804
+ {
1805
+ "epoch": 1.7578659370725034,
1806
+ "grad_norm": 0.9955378174781799,
1807
+ "learning_rate": 1.202235241909333e-05,
1808
+ "loss": 0.1313,
1809
+ "step": 1285
1810
+ },
1811
+ {
1812
+ "epoch": 1.7647058823529411,
1813
+ "grad_norm": 1.0491074323654175,
1814
+ "learning_rate": 1.1911573822695437e-05,
1815
+ "loss": 0.1261,
1816
+ "step": 1290
1817
+ },
1818
+ {
1819
+ "epoch": 1.7715458276333789,
1820
+ "grad_norm": 1.0043137073516846,
1821
+ "learning_rate": 1.180097085535917e-05,
1822
+ "loss": 0.135,
1823
+ "step": 1295
1824
+ },
1825
+ {
1826
+ "epoch": 1.7783857729138166,
1827
+ "grad_norm": 1.1328626871109009,
1828
+ "learning_rate": 1.1690549806726644e-05,
1829
+ "loss": 0.1383,
1830
+ "step": 1300
1831
+ },
1832
+ {
1833
+ "epoch": 1.7852257181942544,
1834
+ "grad_norm": 0.9988858699798584,
1835
+ "learning_rate": 1.1580316956094838e-05,
1836
+ "loss": 0.1238,
1837
+ "step": 1305
1838
+ },
1839
+ {
1840
+ "epoch": 1.792065663474692,
1841
+ "grad_norm": 1.0181727409362793,
1842
+ "learning_rate": 1.1470278572058498e-05,
1843
+ "loss": 0.1283,
1844
+ "step": 1310
1845
+ },
1846
+ {
1847
+ "epoch": 1.7989056087551298,
1848
+ "grad_norm": 0.8925799131393433,
1849
+ "learning_rate": 1.1360440912153672e-05,
1850
+ "loss": 0.1348,
1851
+ "step": 1315
1852
+ },
1853
+ {
1854
+ "epoch": 1.8057455540355676,
1855
+ "grad_norm": 0.9467453360557556,
1856
+ "learning_rate": 1.1250810222501858e-05,
1857
+ "loss": 0.1275,
1858
+ "step": 1320
1859
+ },
1860
+ {
1861
+ "epoch": 1.8125854993160053,
1862
+ "grad_norm": 0.9289389848709106,
1863
+ "learning_rate": 1.1141392737454807e-05,
1864
+ "loss": 0.1309,
1865
+ "step": 1325
1866
+ },
1867
+ {
1868
+ "epoch": 1.819425444596443,
1869
+ "grad_norm": 0.896174430847168,
1870
+ "learning_rate": 1.1032194679239996e-05,
1871
+ "loss": 0.1318,
1872
+ "step": 1330
1873
+ },
1874
+ {
1875
+ "epoch": 1.826265389876881,
1876
+ "grad_norm": 1.178245186805725,
1877
+ "learning_rate": 1.0923222257606795e-05,
1878
+ "loss": 0.1488,
1879
+ "step": 1335
1880
+ },
1881
+ {
1882
+ "epoch": 1.8331053351573188,
1883
+ "grad_norm": 0.996361255645752,
1884
+ "learning_rate": 1.0814481669473318e-05,
1885
+ "loss": 0.1214,
1886
+ "step": 1340
1887
+ },
1888
+ {
1889
+ "epoch": 1.8399452804377565,
1890
+ "grad_norm": 0.845043420791626,
1891
+ "learning_rate": 1.070597909857406e-05,
1892
+ "loss": 0.1297,
1893
+ "step": 1345
1894
+ },
1895
+ {
1896
+ "epoch": 1.8467852257181943,
1897
+ "grad_norm": 0.8970368504524231,
1898
+ "learning_rate": 1.0597720715108216e-05,
1899
+ "loss": 0.1221,
1900
+ "step": 1350
1901
+ },
1902
+ {
1903
+ "epoch": 1.853625170998632,
1904
+ "grad_norm": 0.8815194368362427,
1905
+ "learning_rate": 1.04897126753888e-05,
1906
+ "loss": 0.1077,
1907
+ "step": 1355
1908
+ },
1909
+ {
1910
+ "epoch": 1.8604651162790697,
1911
+ "grad_norm": 0.9306371808052063,
1912
+ "learning_rate": 1.0381961121492588e-05,
1913
+ "loss": 0.1103,
1914
+ "step": 1360
1915
+ },
1916
+ {
1917
+ "epoch": 1.8673050615595075,
1918
+ "grad_norm": 0.8472867012023926,
1919
+ "learning_rate": 1.0274472180910803e-05,
1920
+ "loss": 0.1117,
1921
+ "step": 1365
1922
+ },
1923
+ {
1924
+ "epoch": 1.8741450068399452,
1925
+ "grad_norm": 0.8200199007987976,
1926
+ "learning_rate": 1.0167251966200669e-05,
1927
+ "loss": 0.125,
1928
+ "step": 1370
1929
+ },
1930
+ {
1931
+ "epoch": 1.8809849521203832,
1932
+ "grad_norm": 0.9243358373641968,
1933
+ "learning_rate": 1.0060306574637837e-05,
1934
+ "loss": 0.1205,
1935
+ "step": 1375
1936
+ },
1937
+ {
1938
+ "epoch": 1.887824897400821,
1939
+ "grad_norm": 1.005035638809204,
1940
+ "learning_rate": 9.953642087869613e-06,
1941
+ "loss": 0.1085,
1942
+ "step": 1380
1943
+ },
1944
+ {
1945
+ "epoch": 1.8946648426812587,
1946
+ "grad_norm": 0.8925760984420776,
1947
+ "learning_rate": 9.847264571569144e-06,
1948
+ "loss": 0.1167,
1949
+ "step": 1385
1950
+ },
1951
+ {
1952
+ "epoch": 1.9015047879616964,
1953
+ "grad_norm": 0.8713504672050476,
1954
+ "learning_rate": 9.741180075090468e-06,
1955
+ "loss": 0.1203,
1956
+ "step": 1390
1957
+ },
1958
+ {
1959
+ "epoch": 1.9083447332421342,
1960
+ "grad_norm": 0.7691138982772827,
1961
+ "learning_rate": 9.635394631124506e-06,
1962
+ "loss": 0.1109,
1963
+ "step": 1395
1964
+ },
1965
+ {
1966
+ "epoch": 1.915184678522572,
1967
+ "grad_norm": 0.9341120719909668,
1968
+ "learning_rate": 9.529914255356009e-06,
1969
+ "loss": 0.1148,
1970
+ "step": 1400
1971
+ },
1972
+ {
1973
+ "epoch": 1.9220246238030096,
1974
+ "grad_norm": 0.9697241187095642,
1975
+ "learning_rate": 9.424744946121464e-06,
1976
+ "loss": 0.1081,
1977
+ "step": 1405
1978
+ },
1979
+ {
1980
+ "epoch": 1.9288645690834474,
1981
+ "grad_norm": 1.0418355464935303,
1982
+ "learning_rate": 9.319892684067984e-06,
1983
+ "loss": 0.1195,
1984
+ "step": 1410
1985
+ },
1986
+ {
1987
+ "epoch": 1.9357045143638851,
1988
+ "grad_norm": 0.992321252822876,
1989
+ "learning_rate": 9.215363431813196e-06,
1990
+ "loss": 0.1052,
1991
+ "step": 1415
1992
+ },
1993
+ {
1994
+ "epoch": 1.9425444596443229,
1995
+ "grad_norm": 1.0079401731491089,
1996
+ "learning_rate": 9.111163133606195e-06,
1997
+ "loss": 0.1123,
1998
+ "step": 1420
1999
+ },
2000
+ {
2001
+ "epoch": 1.9493844049247606,
2002
+ "grad_norm": 0.8518632650375366,
2003
+ "learning_rate": 9.007297714989485e-06,
2004
+ "loss": 0.1057,
2005
+ "step": 1425
2006
+ },
2007
+ {
2008
+ "epoch": 1.9562243502051984,
2009
+ "grad_norm": 1.0242724418640137,
2010
+ "learning_rate": 8.903773082462026e-06,
2011
+ "loss": 0.1102,
2012
+ "step": 1430
2013
+ },
2014
+ {
2015
+ "epoch": 1.963064295485636,
2016
+ "grad_norm": 0.8795663118362427,
2017
+ "learning_rate": 8.800595123143352e-06,
2018
+ "loss": 0.0979,
2019
+ "step": 1435
2020
+ },
2021
+ {
2022
+ "epoch": 1.9699042407660738,
2023
+ "grad_norm": 0.7944589853286743,
2024
+ "learning_rate": 8.697769704438778e-06,
2025
+ "loss": 0.0994,
2026
+ "step": 1440
2027
+ },
2028
+ {
2029
+ "epoch": 1.9767441860465116,
2030
+ "grad_norm": 0.9179325103759766,
2031
+ "learning_rate": 8.59530267370575e-06,
2032
+ "loss": 0.1032,
2033
+ "step": 1445
2034
+ },
2035
+ {
2036
+ "epoch": 1.9835841313269493,
2037
+ "grad_norm": 0.9672313928604126,
2038
+ "learning_rate": 8.493199857921311e-06,
2039
+ "loss": 0.1047,
2040
+ "step": 1450
2041
+ },
2042
+ {
2043
+ "epoch": 1.990424076607387,
2044
+ "grad_norm": 0.8533863425254822,
2045
+ "learning_rate": 8.39146706335076e-06,
2046
+ "loss": 0.0944,
2047
+ "step": 1455
2048
+ },
2049
+ {
2050
+ "epoch": 1.9972640218878248,
2051
+ "grad_norm": 0.7924801707267761,
2052
+ "learning_rate": 8.29011007521745e-06,
2053
+ "loss": 0.1084,
2054
+ "step": 1460
2055
+ },
2056
+ {
2057
+ "epoch": 2.0041039671682626,
2058
+ "grad_norm": 0.6392746567726135,
2059
+ "learning_rate": 8.189134657373809e-06,
2060
+ "loss": 0.0833,
2061
+ "step": 1465
2062
+ },
2063
+ {
2064
+ "epoch": 2.0109439124487003,
2065
+ "grad_norm": 0.71649569272995,
2066
+ "learning_rate": 8.088546551973562e-06,
2067
+ "loss": 0.0748,
2068
+ "step": 1470
2069
+ },
2070
+ {
2071
+ "epoch": 2.017783857729138,
2072
+ "grad_norm": 0.823631227016449,
2073
+ "learning_rate": 7.988351479145202e-06,
2074
+ "loss": 0.0736,
2075
+ "step": 1475
2076
+ },
2077
+ {
2078
+ "epoch": 2.024623803009576,
2079
+ "grad_norm": 0.747878909111023,
2080
+ "learning_rate": 7.888555136666684e-06,
2081
+ "loss": 0.0865,
2082
+ "step": 1480
2083
+ },
2084
+ {
2085
+ "epoch": 2.0314637482900135,
2086
+ "grad_norm": 0.6734447479248047,
2087
+ "learning_rate": 7.78916319964144e-06,
2088
+ "loss": 0.0921,
2089
+ "step": 1485
2090
+ },
2091
+ {
2092
+ "epoch": 2.0383036935704513,
2093
+ "grad_norm": 0.6489822864532471,
2094
+ "learning_rate": 7.690181320175636e-06,
2095
+ "loss": 0.08,
2096
+ "step": 1490
2097
+ },
2098
+ {
2099
+ "epoch": 2.045143638850889,
2100
+ "grad_norm": 0.7271413803100586,
2101
+ "learning_rate": 7.591615127056758e-06,
2102
+ "loss": 0.0819,
2103
+ "step": 1495
2104
+ },
2105
+ {
2106
+ "epoch": 2.0519835841313268,
2107
+ "grad_norm": 0.9902181625366211,
2108
+ "learning_rate": 7.493470225433507e-06,
2109
+ "loss": 0.0787,
2110
+ "step": 1500
2111
+ },
2112
+ {
2113
+ "epoch": 2.0588235294117645,
2114
+ "grad_norm": 0.6948493719100952,
2115
+ "learning_rate": 7.395752196497077e-06,
2116
+ "loss": 0.0807,
2117
+ "step": 1505
2118
+ },
2119
+ {
2120
+ "epoch": 2.0656634746922027,
2121
+ "grad_norm": 0.670220136642456,
2122
+ "learning_rate": 7.298466597163766e-06,
2123
+ "loss": 0.0847,
2124
+ "step": 1510
2125
+ },
2126
+ {
2127
+ "epoch": 2.0725034199726404,
2128
+ "grad_norm": 0.7845743894577026,
2129
+ "learning_rate": 7.201618959758935e-06,
2130
+ "loss": 0.0787,
2131
+ "step": 1515
2132
+ },
2133
+ {
2134
+ "epoch": 2.079343365253078,
2135
+ "grad_norm": 0.7086281776428223,
2136
+ "learning_rate": 7.1052147917024564e-06,
2137
+ "loss": 0.0898,
2138
+ "step": 1520
2139
+ },
2140
+ {
2141
+ "epoch": 2.086183310533516,
2142
+ "grad_norm": 0.7347419857978821,
2143
+ "learning_rate": 7.009259575195492e-06,
2144
+ "loss": 0.0805,
2145
+ "step": 1525
2146
+ },
2147
+ {
2148
+ "epoch": 2.0930232558139537,
2149
+ "grad_norm": 0.7243831157684326,
2150
+ "learning_rate": 6.9137587669087405e-06,
2151
+ "loss": 0.0851,
2152
+ "step": 1530
2153
+ },
2154
+ {
2155
+ "epoch": 2.0998632010943914,
2156
+ "grad_norm": 0.8953825235366821,
2157
+ "learning_rate": 6.818717797672141e-06,
2158
+ "loss": 0.0905,
2159
+ "step": 1535
2160
+ },
2161
+ {
2162
+ "epoch": 2.106703146374829,
2163
+ "grad_norm": 0.6968517899513245,
2164
+ "learning_rate": 6.724142072166038e-06,
2165
+ "loss": 0.079,
2166
+ "step": 1540
2167
+ },
2168
+ {
2169
+ "epoch": 2.113543091655267,
2170
+ "grad_norm": 0.6811033487319946,
2171
+ "learning_rate": 6.630036968613816e-06,
2172
+ "loss": 0.0785,
2173
+ "step": 1545
2174
+ },
2175
+ {
2176
+ "epoch": 2.1203830369357046,
2177
+ "grad_norm": 0.7520840764045715,
2178
+ "learning_rate": 6.536407838476091e-06,
2179
+ "loss": 0.0753,
2180
+ "step": 1550
2181
+ },
2182
+ {
2183
+ "epoch": 2.1272229822161424,
2184
+ "grad_norm": 0.6734663844108582,
2185
+ "learning_rate": 6.443260006146368e-06,
2186
+ "loss": 0.0772,
2187
+ "step": 1555
2188
+ },
2189
+ {
2190
+ "epoch": 2.13406292749658,
2191
+ "grad_norm": 0.7304068803787231,
2192
+ "learning_rate": 6.350598768648248e-06,
2193
+ "loss": 0.0731,
2194
+ "step": 1560
2195
+ },
2196
+ {
2197
+ "epoch": 2.140902872777018,
2198
+ "grad_norm": 0.80950528383255,
2199
+ "learning_rate": 6.258429395334237e-06,
2200
+ "loss": 0.0692,
2201
+ "step": 1565
2202
+ },
2203
+ {
2204
+ "epoch": 2.1477428180574556,
2205
+ "grad_norm": 0.7778096199035645,
2206
+ "learning_rate": 6.166757127586067e-06,
2207
+ "loss": 0.0722,
2208
+ "step": 1570
2209
+ },
2210
+ {
2211
+ "epoch": 2.1545827633378933,
2212
+ "grad_norm": 0.7647256255149841,
2213
+ "learning_rate": 6.075587178516635e-06,
2214
+ "loss": 0.0834,
2215
+ "step": 1575
2216
+ },
2217
+ {
2218
+ "epoch": 2.161422708618331,
2219
+ "grad_norm": 0.7058929800987244,
2220
+ "learning_rate": 5.984924732673565e-06,
2221
+ "loss": 0.0845,
2222
+ "step": 1580
2223
+ },
2224
+ {
2225
+ "epoch": 2.168262653898769,
2226
+ "grad_norm": 0.6837173104286194,
2227
+ "learning_rate": 5.894774945744392e-06,
2228
+ "loss": 0.0779,
2229
+ "step": 1585
2230
+ },
2231
+ {
2232
+ "epoch": 2.1751025991792066,
2233
+ "grad_norm": 0.7722723484039307,
2234
+ "learning_rate": 5.805142944263319e-06,
2235
+ "loss": 0.0794,
2236
+ "step": 1590
2237
+ },
2238
+ {
2239
+ "epoch": 2.1819425444596443,
2240
+ "grad_norm": 0.8545292019844055,
2241
+ "learning_rate": 5.716033825319745e-06,
2242
+ "loss": 0.079,
2243
+ "step": 1595
2244
+ },
2245
+ {
2246
+ "epoch": 2.188782489740082,
2247
+ "grad_norm": 0.6731324791908264,
2248
+ "learning_rate": 5.627452656268387e-06,
2249
+ "loss": 0.0776,
2250
+ "step": 1600
2251
+ },
2252
+ {
2253
+ "epoch": 2.19562243502052,
2254
+ "grad_norm": 0.7811100482940674,
2255
+ "learning_rate": 5.539404474441098e-06,
2256
+ "loss": 0.08,
2257
+ "step": 1605
2258
+ },
2259
+ {
2260
+ "epoch": 2.2024623803009575,
2261
+ "grad_norm": 0.7762587070465088,
2262
+ "learning_rate": 5.45189428686045e-06,
2263
+ "loss": 0.083,
2264
+ "step": 1610
2265
+ },
2266
+ {
2267
+ "epoch": 2.2093023255813953,
2268
+ "grad_norm": 0.7523200511932373,
2269
+ "learning_rate": 5.364927069954971e-06,
2270
+ "loss": 0.0871,
2271
+ "step": 1615
2272
+ },
2273
+ {
2274
+ "epoch": 2.216142270861833,
2275
+ "grad_norm": 0.7899711728096008,
2276
+ "learning_rate": 5.2785077692761455e-06,
2277
+ "loss": 0.079,
2278
+ "step": 1620
2279
+ },
2280
+ {
2281
+ "epoch": 2.2229822161422708,
2282
+ "grad_norm": 0.8483776450157166,
2283
+ "learning_rate": 5.192641299217206e-06,
2284
+ "loss": 0.078,
2285
+ "step": 1625
2286
+ },
2287
+ {
2288
+ "epoch": 2.2298221614227085,
2289
+ "grad_norm": 0.6784462332725525,
2290
+ "learning_rate": 5.107332542733647e-06,
2291
+ "loss": 0.074,
2292
+ "step": 1630
2293
+ },
2294
+ {
2295
+ "epoch": 2.2366621067031462,
2296
+ "grad_norm": 0.7743968367576599,
2297
+ "learning_rate": 5.022586351065536e-06,
2298
+ "loss": 0.0776,
2299
+ "step": 1635
2300
+ },
2301
+ {
2302
+ "epoch": 2.243502051983584,
2303
+ "grad_norm": 0.7630516886711121,
2304
+ "learning_rate": 4.938407543461665e-06,
2305
+ "loss": 0.0733,
2306
+ "step": 1640
2307
+ },
2308
+ {
2309
+ "epoch": 2.2503419972640217,
2310
+ "grad_norm": 0.6820114850997925,
2311
+ "learning_rate": 4.8548009069054945e-06,
2312
+ "loss": 0.0718,
2313
+ "step": 1645
2314
+ },
2315
+ {
2316
+ "epoch": 2.2571819425444595,
2317
+ "grad_norm": 0.7109975218772888,
2318
+ "learning_rate": 4.771771195842895e-06,
2319
+ "loss": 0.0756,
2320
+ "step": 1650
2321
+ },
2322
+ {
2323
+ "epoch": 2.264021887824897,
2324
+ "grad_norm": 0.6362276673316956,
2325
+ "learning_rate": 4.689323131911818e-06,
2326
+ "loss": 0.0746,
2327
+ "step": 1655
2328
+ },
2329
+ {
2330
+ "epoch": 2.270861833105335,
2331
+ "grad_norm": 0.6837899684906006,
2332
+ "learning_rate": 4.6074614036737785e-06,
2333
+ "loss": 0.0782,
2334
+ "step": 1660
2335
+ },
2336
+ {
2337
+ "epoch": 2.277701778385773,
2338
+ "grad_norm": 0.7597292065620422,
2339
+ "learning_rate": 4.5261906663472155e-06,
2340
+ "loss": 0.0726,
2341
+ "step": 1665
2342
+ },
2343
+ {
2344
+ "epoch": 2.2845417236662104,
2345
+ "grad_norm": 0.6519300937652588,
2346
+ "learning_rate": 4.445515541542788e-06,
2347
+ "loss": 0.0734,
2348
+ "step": 1670
2349
+ },
2350
+ {
2351
+ "epoch": 2.2913816689466486,
2352
+ "grad_norm": 0.6085669994354248,
2353
+ "learning_rate": 4.365440617000555e-06,
2354
+ "loss": 0.077,
2355
+ "step": 1675
2356
+ },
2357
+ {
2358
+ "epoch": 2.2982216142270864,
2359
+ "grad_norm": 0.7052056193351746,
2360
+ "learning_rate": 4.285970446329055e-06,
2361
+ "loss": 0.0735,
2362
+ "step": 1680
2363
+ },
2364
+ {
2365
+ "epoch": 2.305061559507524,
2366
+ "grad_norm": 0.5915761590003967,
2367
+ "learning_rate": 4.207109548746404e-06,
2368
+ "loss": 0.0713,
2369
+ "step": 1685
2370
+ },
2371
+ {
2372
+ "epoch": 2.311901504787962,
2373
+ "grad_norm": 0.8734668493270874,
2374
+ "learning_rate": 4.128862408823268e-06,
2375
+ "loss": 0.0855,
2376
+ "step": 1690
2377
+ },
2378
+ {
2379
+ "epoch": 2.3187414500683996,
2380
+ "grad_norm": 0.6716635227203369,
2381
+ "learning_rate": 4.051233476227853e-06,
2382
+ "loss": 0.0699,
2383
+ "step": 1695
2384
+ },
2385
+ {
2386
+ "epoch": 2.3255813953488373,
2387
+ "grad_norm": 0.5665057897567749,
2388
+ "learning_rate": 3.974227165472853e-06,
2389
+ "loss": 0.0635,
2390
+ "step": 1700
2391
+ },
2392
+ {
2393
+ "epoch": 2.332421340629275,
2394
+ "grad_norm": 0.5990927815437317,
2395
+ "learning_rate": 3.897847855664432e-06,
2396
+ "loss": 0.0708,
2397
+ "step": 1705
2398
+ },
2399
+ {
2400
+ "epoch": 2.339261285909713,
2401
+ "grad_norm": 0.8617225289344788,
2402
+ "learning_rate": 3.822099890253181e-06,
2403
+ "loss": 0.0616,
2404
+ "step": 1710
2405
+ },
2406
+ {
2407
+ "epoch": 2.3461012311901506,
2408
+ "grad_norm": 0.5008922815322876,
2409
+ "learning_rate": 3.7469875767871288e-06,
2410
+ "loss": 0.0628,
2411
+ "step": 1715
2412
+ },
2413
+ {
2414
+ "epoch": 2.3529411764705883,
2415
+ "grad_norm": 0.7114216685295105,
2416
+ "learning_rate": 3.672515186666781e-06,
2417
+ "loss": 0.0663,
2418
+ "step": 1720
2419
+ },
2420
+ {
2421
+ "epoch": 2.359781121751026,
2422
+ "grad_norm": 0.6179472208023071,
2423
+ "learning_rate": 3.598686954902211e-06,
2424
+ "loss": 0.074,
2425
+ "step": 1725
2426
+ },
2427
+ {
2428
+ "epoch": 2.366621067031464,
2429
+ "grad_norm": 0.7750431299209595,
2430
+ "learning_rate": 3.52550707987225e-06,
2431
+ "loss": 0.0723,
2432
+ "step": 1730
2433
+ },
2434
+ {
2435
+ "epoch": 2.3734610123119015,
2436
+ "grad_norm": 0.5394264459609985,
2437
+ "learning_rate": 3.452979723085717e-06,
2438
+ "loss": 0.0681,
2439
+ "step": 1735
2440
+ },
2441
+ {
2442
+ "epoch": 2.3803009575923393,
2443
+ "grad_norm": 0.6713101267814636,
2444
+ "learning_rate": 3.3811090089447816e-06,
2445
+ "loss": 0.0715,
2446
+ "step": 1740
2447
+ },
2448
+ {
2449
+ "epoch": 2.387140902872777,
2450
+ "grad_norm": 0.7275838255882263,
2451
+ "learning_rate": 3.3098990245104057e-06,
2452
+ "loss": 0.0642,
2453
+ "step": 1745
2454
+ },
2455
+ {
2456
+ "epoch": 2.3939808481532148,
2457
+ "grad_norm": 0.6464980244636536,
2458
+ "learning_rate": 3.2393538192699457e-06,
2459
+ "loss": 0.074,
2460
+ "step": 1750
2461
+ },
2462
+ {
2463
+ "epoch": 2.4008207934336525,
2464
+ "grad_norm": 0.610549807548523,
2465
+ "learning_rate": 3.1694774049068594e-06,
2466
+ "loss": 0.0616,
2467
+ "step": 1755
2468
+ },
2469
+ {
2470
+ "epoch": 2.4076607387140903,
2471
+ "grad_norm": 0.6246511936187744,
2472
+ "learning_rate": 3.1002737550725674e-06,
2473
+ "loss": 0.0673,
2474
+ "step": 1760
2475
+ },
2476
+ {
2477
+ "epoch": 2.414500683994528,
2478
+ "grad_norm": 0.5291991829872131,
2479
+ "learning_rate": 3.0317468051604996e-06,
2480
+ "loss": 0.0594,
2481
+ "step": 1765
2482
+ },
2483
+ {
2484
+ "epoch": 2.4213406292749657,
2485
+ "grad_norm": 0.6580734848976135,
2486
+ "learning_rate": 2.9639004520822927e-06,
2487
+ "loss": 0.0698,
2488
+ "step": 1770
2489
+ },
2490
+ {
2491
+ "epoch": 2.4281805745554035,
2492
+ "grad_norm": 0.6327461004257202,
2493
+ "learning_rate": 2.896738554046189e-06,
2494
+ "loss": 0.0672,
2495
+ "step": 1775
2496
+ },
2497
+ {
2498
+ "epoch": 2.4350205198358412,
2499
+ "grad_norm": 0.527522087097168,
2500
+ "learning_rate": 2.8302649303376242e-06,
2501
+ "loss": 0.0685,
2502
+ "step": 1780
2503
+ },
2504
+ {
2505
+ "epoch": 2.441860465116279,
2506
+ "grad_norm": 0.7473851442337036,
2507
+ "learning_rate": 2.7644833611020544e-06,
2508
+ "loss": 0.0741,
2509
+ "step": 1785
2510
+ },
2511
+ {
2512
+ "epoch": 2.4487004103967167,
2513
+ "grad_norm": 0.6569819450378418,
2514
+ "learning_rate": 2.699397587129958e-06,
2515
+ "loss": 0.0692,
2516
+ "step": 1790
2517
+ },
2518
+ {
2519
+ "epoch": 2.4555403556771545,
2520
+ "grad_norm": 0.5640328526496887,
2521
+ "learning_rate": 2.6350113096441508e-06,
2522
+ "loss": 0.0726,
2523
+ "step": 1795
2524
+ },
2525
+ {
2526
+ "epoch": 2.462380300957592,
2527
+ "grad_norm": 0.6890896558761597,
2528
+ "learning_rate": 2.5713281900892818e-06,
2529
+ "loss": 0.0739,
2530
+ "step": 1800
2531
+ },
2532
+ {
2533
+ "epoch": 2.46922024623803,
2534
+ "grad_norm": 0.535921573638916,
2535
+ "learning_rate": 2.5083518499236167e-06,
2536
+ "loss": 0.0783,
2537
+ "step": 1805
2538
+ },
2539
+ {
2540
+ "epoch": 2.4760601915184677,
2541
+ "grad_norm": 0.6307790875434875,
2542
+ "learning_rate": 2.4460858704131174e-06,
2543
+ "loss": 0.0646,
2544
+ "step": 1810
2545
+ },
2546
+ {
2547
+ "epoch": 2.4829001367989054,
2548
+ "grad_norm": 0.6907503008842468,
2549
+ "learning_rate": 2.384533792427768e-06,
2550
+ "loss": 0.0635,
2551
+ "step": 1815
2552
+ },
2553
+ {
2554
+ "epoch": 2.4897400820793436,
2555
+ "grad_norm": 0.5222296714782715,
2556
+ "learning_rate": 2.3236991162402216e-06,
2557
+ "loss": 0.064,
2558
+ "step": 1820
2559
+ },
2560
+ {
2561
+ "epoch": 2.496580027359781,
2562
+ "grad_norm": 0.6377049684524536,
2563
+ "learning_rate": 2.263585301326754e-06,
2564
+ "loss": 0.0732,
2565
+ "step": 1825
2566
+ },
2567
+ {
2568
+ "epoch": 2.503419972640219,
2569
+ "grad_norm": 0.5305001735687256,
2570
+ "learning_rate": 2.204195766170543e-06,
2571
+ "loss": 0.064,
2572
+ "step": 1830
2573
+ },
2574
+ {
2575
+ "epoch": 2.5102599179206564,
2576
+ "grad_norm": 0.5369933247566223,
2577
+ "learning_rate": 2.1455338880672393e-06,
2578
+ "loss": 0.0789,
2579
+ "step": 1835
2580
+ },
2581
+ {
2582
+ "epoch": 2.5170998632010946,
2583
+ "grad_norm": 0.6566374897956848,
2584
+ "learning_rate": 2.0876030029329477e-06,
2585
+ "loss": 0.0664,
2586
+ "step": 1840
2587
+ },
2588
+ {
2589
+ "epoch": 2.523939808481532,
2590
+ "grad_norm": 0.5830959677696228,
2591
+ "learning_rate": 2.030406405114501e-06,
2592
+ "loss": 0.0644,
2593
+ "step": 1845
2594
+ },
2595
+ {
2596
+ "epoch": 2.53077975376197,
2597
+ "grad_norm": 0.5287669897079468,
2598
+ "learning_rate": 1.9739473472021226e-06,
2599
+ "loss": 0.067,
2600
+ "step": 1850
2601
+ },
2602
+ {
2603
+ "epoch": 2.537619699042408,
2604
+ "grad_norm": 0.6011587381362915,
2605
+ "learning_rate": 1.9182290398444725e-06,
2606
+ "loss": 0.0761,
2607
+ "step": 1855
2608
+ },
2609
+ {
2610
+ "epoch": 2.5444596443228455,
2611
+ "grad_norm": 0.5890557169914246,
2612
+ "learning_rate": 1.8632546515660621e-06,
2613
+ "loss": 0.056,
2614
+ "step": 1860
2615
+ },
2616
+ {
2617
+ "epoch": 2.5512995896032833,
2618
+ "grad_norm": 0.45841431617736816,
2619
+ "learning_rate": 1.8090273085870624e-06,
2620
+ "loss": 0.0621,
2621
+ "step": 1865
2622
+ },
2623
+ {
2624
+ "epoch": 2.558139534883721,
2625
+ "grad_norm": 0.5593291521072388,
2626
+ "learning_rate": 1.7555500946455399e-06,
2627
+ "loss": 0.0612,
2628
+ "step": 1870
2629
+ },
2630
+ {
2631
+ "epoch": 2.5649794801641588,
2632
+ "grad_norm": 0.7063523530960083,
2633
+ "learning_rate": 1.702826050822089e-06,
2634
+ "loss": 0.0698,
2635
+ "step": 1875
2636
+ },
2637
+ {
2638
+ "epoch": 2.5718194254445965,
2639
+ "grad_norm": 0.6223360896110535,
2640
+ "learning_rate": 1.6508581753668823e-06,
2641
+ "loss": 0.0711,
2642
+ "step": 1880
2643
+ },
2644
+ {
2645
+ "epoch": 2.5786593707250343,
2646
+ "grad_norm": 0.5816681385040283,
2647
+ "learning_rate": 1.5996494235291903e-06,
2648
+ "loss": 0.0662,
2649
+ "step": 1885
2650
+ },
2651
+ {
2652
+ "epoch": 2.585499316005472,
2653
+ "grad_norm": 0.562164843082428,
2654
+ "learning_rate": 1.54920270738932e-06,
2655
+ "loss": 0.0714,
2656
+ "step": 1890
2657
+ },
2658
+ {
2659
+ "epoch": 2.5923392612859097,
2660
+ "grad_norm": 0.5410878658294678,
2661
+ "learning_rate": 1.4995208956929957e-06,
2662
+ "loss": 0.0707,
2663
+ "step": 1895
2664
+ },
2665
+ {
2666
+ "epoch": 2.5991792065663475,
2667
+ "grad_norm": 0.5669549703598022,
2668
+ "learning_rate": 1.4506068136882528e-06,
2669
+ "loss": 0.0755,
2670
+ "step": 1900
2671
+ },
2672
+ {
2673
+ "epoch": 2.6060191518467852,
2674
+ "grad_norm": 0.5971465110778809,
2675
+ "learning_rate": 1.40246324296475e-06,
2676
+ "loss": 0.0721,
2677
+ "step": 1905
2678
+ },
2679
+ {
2680
+ "epoch": 2.612859097127223,
2681
+ "grad_norm": 0.6834708452224731,
2682
+ "learning_rate": 1.3550929212955971e-06,
2683
+ "loss": 0.0699,
2684
+ "step": 1910
2685
+ },
2686
+ {
2687
+ "epoch": 2.6196990424076607,
2688
+ "grad_norm": 0.4914969503879547,
2689
+ "learning_rate": 1.3084985424816715e-06,
2690
+ "loss": 0.0618,
2691
+ "step": 1915
2692
+ },
2693
+ {
2694
+ "epoch": 2.6265389876880985,
2695
+ "grad_norm": 0.6790369749069214,
2696
+ "learning_rate": 1.2626827561984305e-06,
2697
+ "loss": 0.063,
2698
+ "step": 1920
2699
+ },
2700
+ {
2701
+ "epoch": 2.633378932968536,
2702
+ "grad_norm": 0.5151853561401367,
2703
+ "learning_rate": 1.2176481678452172e-06,
2704
+ "loss": 0.0706,
2705
+ "step": 1925
2706
+ },
2707
+ {
2708
+ "epoch": 2.640218878248974,
2709
+ "grad_norm": 0.5760762691497803,
2710
+ "learning_rate": 1.1733973383971186e-06,
2711
+ "loss": 0.0632,
2712
+ "step": 1930
2713
+ },
2714
+ {
2715
+ "epoch": 2.6470588235294117,
2716
+ "grad_norm": 0.5759361386299133,
2717
+ "learning_rate": 1.1299327842593237e-06,
2718
+ "loss": 0.0609,
2719
+ "step": 1935
2720
+ },
2721
+ {
2722
+ "epoch": 2.6538987688098494,
2723
+ "grad_norm": 0.6918815970420837,
2724
+ "learning_rate": 1.0872569771240203e-06,
2725
+ "loss": 0.0655,
2726
+ "step": 1940
2727
+ },
2728
+ {
2729
+ "epoch": 2.660738714090287,
2730
+ "grad_norm": 0.5681085586547852,
2731
+ "learning_rate": 1.0453723438298368e-06,
2732
+ "loss": 0.0672,
2733
+ "step": 1945
2734
+ },
2735
+ {
2736
+ "epoch": 2.667578659370725,
2737
+ "grad_norm": 0.5260602831840515,
2738
+ "learning_rate": 1.0042812662238527e-06,
2739
+ "loss": 0.0625,
2740
+ "step": 1950
2741
+ },
2742
+ {
2743
+ "epoch": 2.6744186046511627,
2744
+ "grad_norm": 0.6265805959701538,
2745
+ "learning_rate": 9.639860810261187e-07,
2746
+ "loss": 0.0668,
2747
+ "step": 1955
2748
+ },
2749
+ {
2750
+ "epoch": 2.6812585499316004,
2751
+ "grad_norm": 0.6276566386222839,
2752
+ "learning_rate": 9.244890796968048e-07,
2753
+ "loss": 0.0626,
2754
+ "step": 1960
2755
+ },
2756
+ {
2757
+ "epoch": 2.688098495212038,
2758
+ "grad_norm": 0.5453227758407593,
2759
+ "learning_rate": 8.857925083058794e-07,
2760
+ "loss": 0.0718,
2761
+ "step": 1965
2762
+ },
2763
+ {
2764
+ "epoch": 2.694938440492476,
2765
+ "grad_norm": 0.5649544596672058,
2766
+ "learning_rate": 8.478985674053797e-07,
2767
+ "loss": 0.06,
2768
+ "step": 1970
2769
+ },
2770
+ {
2771
+ "epoch": 2.701778385772914,
2772
+ "grad_norm": 0.5511782765388489,
2773
+ "learning_rate": 8.108094119042769e-07,
2774
+ "loss": 0.0682,
2775
+ "step": 1975
2776
+ },
2777
+ {
2778
+ "epoch": 2.7086183310533514,
2779
+ "grad_norm": 0.5780085921287537,
2780
+ "learning_rate": 7.745271509459379e-07,
2781
+ "loss": 0.0602,
2782
+ "step": 1980
2783
+ },
2784
+ {
2785
+ "epoch": 2.7154582763337896,
2786
+ "grad_norm": 0.6196208000183105,
2787
+ "learning_rate": 7.39053847788177e-07,
2788
+ "loss": 0.0629,
2789
+ "step": 1985
2790
+ },
2791
+ {
2792
+ "epoch": 2.722298221614227,
2793
+ "grad_norm": 0.5995792150497437,
2794
+ "learning_rate": 7.043915196859274e-07,
2795
+ "loss": 0.0696,
2796
+ "step": 1990
2797
+ },
2798
+ {
2799
+ "epoch": 2.729138166894665,
2800
+ "grad_norm": 0.5502211451530457,
2801
+ "learning_rate": 6.705421377765276e-07,
2802
+ "loss": 0.0652,
2803
+ "step": 1995
2804
+ },
2805
+ {
2806
+ "epoch": 2.7359781121751023,
2807
+ "grad_norm": 0.6088355779647827,
2808
+ "learning_rate": 6.375076269676356e-07,
2809
+ "loss": 0.0649,
2810
+ "step": 2000
2811
+ },
2812
+ {
2813
+ "epoch": 2.7428180574555405,
2814
+ "grad_norm": 0.5751022100448608,
2815
+ "learning_rate": 6.052898658277439e-07,
2816
+ "loss": 0.0674,
2817
+ "step": 2005
2818
+ },
2819
+ {
2820
+ "epoch": 2.7496580027359783,
2821
+ "grad_norm": 0.5008280873298645,
2822
+ "learning_rate": 5.73890686479378e-07,
2823
+ "loss": 0.0632,
2824
+ "step": 2010
2825
+ },
2826
+ {
2827
+ "epoch": 2.756497948016416,
2828
+ "grad_norm": 0.4570299983024597,
2829
+ "learning_rate": 5.433118744948979e-07,
2830
+ "loss": 0.0556,
2831
+ "step": 2015
2832
+ },
2833
+ {
2834
+ "epoch": 2.7633378932968538,
2835
+ "grad_norm": 0.48228806257247925,
2836
+ "learning_rate": 5.135551687949475e-07,
2837
+ "loss": 0.0613,
2838
+ "step": 2020
2839
+ },
2840
+ {
2841
+ "epoch": 2.7701778385772915,
2842
+ "grad_norm": 0.6606765389442444,
2843
+ "learning_rate": 4.84622261549586e-07,
2844
+ "loss": 0.0643,
2845
+ "step": 2025
2846
+ },
2847
+ {
2848
+ "epoch": 2.7770177838577292,
2849
+ "grad_norm": 0.6221924424171448,
2850
+ "learning_rate": 4.565147980820494e-07,
2851
+ "loss": 0.0659,
2852
+ "step": 2030
2853
+ },
2854
+ {
2855
+ "epoch": 2.783857729138167,
2856
+ "grad_norm": 0.5815507173538208,
2857
+ "learning_rate": 4.2923437677518383e-07,
2858
+ "loss": 0.0608,
2859
+ "step": 2035
2860
+ },
2861
+ {
2862
+ "epoch": 2.7906976744186047,
2863
+ "grad_norm": 0.4951245188713074,
2864
+ "learning_rate": 4.027825489805598e-07,
2865
+ "loss": 0.0695,
2866
+ "step": 2040
2867
+ },
2868
+ {
2869
+ "epoch": 2.7975376196990425,
2870
+ "grad_norm": 0.6763774156570435,
2871
+ "learning_rate": 3.771608189302461e-07,
2872
+ "loss": 0.0588,
2873
+ "step": 2045
2874
+ },
2875
+ {
2876
+ "epoch": 2.80437756497948,
2877
+ "grad_norm": 0.5480711460113525,
2878
+ "learning_rate": 3.5237064365126524e-07,
2879
+ "loss": 0.0578,
2880
+ "step": 2050
2881
+ },
2882
+ {
2883
+ "epoch": 2.811217510259918,
2884
+ "grad_norm": 0.4940590560436249,
2885
+ "learning_rate": 3.284134328827443e-07,
2886
+ "loss": 0.0613,
2887
+ "step": 2055
2888
+ },
2889
+ {
2890
+ "epoch": 2.8180574555403557,
2891
+ "grad_norm": 0.49339720606803894,
2892
+ "learning_rate": 3.052905489957475e-07,
2893
+ "loss": 0.0649,
2894
+ "step": 2060
2895
+ },
2896
+ {
2897
+ "epoch": 2.8248974008207934,
2898
+ "grad_norm": 0.5984700918197632,
2899
+ "learning_rate": 2.8300330691579e-07,
2900
+ "loss": 0.0634,
2901
+ "step": 2065
2902
+ },
2903
+ {
2904
+ "epoch": 2.831737346101231,
2905
+ "grad_norm": 0.4606676697731018,
2906
+ "learning_rate": 2.6155297404807933e-07,
2907
+ "loss": 0.0535,
2908
+ "step": 2070
2909
+ },
2910
+ {
2911
+ "epoch": 2.838577291381669,
2912
+ "grad_norm": 0.6329459547996521,
2913
+ "learning_rate": 2.4094077020543283e-07,
2914
+ "loss": 0.0662,
2915
+ "step": 2075
2916
+ },
2917
+ {
2918
+ "epoch": 2.8454172366621067,
2919
+ "grad_norm": 0.5132828950881958,
2920
+ "learning_rate": 2.2116786753891005e-07,
2921
+ "loss": 0.0561,
2922
+ "step": 2080
2923
+ },
2924
+ {
2925
+ "epoch": 2.8522571819425444,
2926
+ "grad_norm": 0.5690356492996216,
2927
+ "learning_rate": 2.0223539047115935e-07,
2928
+ "loss": 0.0566,
2929
+ "step": 2085
2930
+ },
2931
+ {
2932
+ "epoch": 2.859097127222982,
2933
+ "grad_norm": 0.5672386884689331,
2934
+ "learning_rate": 1.8414441563247398e-07,
2935
+ "loss": 0.0636,
2936
+ "step": 2090
2937
+ },
2938
+ {
2939
+ "epoch": 2.86593707250342,
2940
+ "grad_norm": 0.5576285719871521,
2941
+ "learning_rate": 1.6689597179956772e-07,
2942
+ "loss": 0.0668,
2943
+ "step": 2095
2944
+ },
2945
+ {
2946
+ "epoch": 2.8727770177838576,
2947
+ "grad_norm": 0.6511788368225098,
2948
+ "learning_rate": 1.5049103983707346e-07,
2949
+ "loss": 0.0681,
2950
+ "step": 2100
2951
+ },
2952
+ {
2953
+ "epoch": 2.8796169630642954,
2954
+ "grad_norm": 0.5570366382598877,
2955
+ "learning_rate": 1.3493055264176102e-07,
2956
+ "loss": 0.0586,
2957
+ "step": 2105
2958
+ },
2959
+ {
2960
+ "epoch": 2.886456908344733,
2961
+ "grad_norm": 0.5808674097061157,
2962
+ "learning_rate": 1.202153950894863e-07,
2963
+ "loss": 0.0616,
2964
+ "step": 2110
2965
+ },
2966
+ {
2967
+ "epoch": 2.893296853625171,
2968
+ "grad_norm": 0.6499027609825134,
2969
+ "learning_rate": 1.0634640398487993e-07,
2970
+ "loss": 0.0668,
2971
+ "step": 2115
2972
+ },
2973
+ {
2974
+ "epoch": 2.9001367989056086,
2975
+ "grad_norm": 0.5679591298103333,
2976
+ "learning_rate": 9.332436801374522e-08,
2977
+ "loss": 0.061,
2978
+ "step": 2120
2979
+ },
2980
+ {
2981
+ "epoch": 2.9069767441860463,
2982
+ "grad_norm": 0.4444182813167572,
2983
+ "learning_rate": 8.115002769822578e-08,
2984
+ "loss": 0.0549,
2985
+ "step": 2125
2986
+ },
2987
+ {
2988
+ "epoch": 2.9138166894664845,
2989
+ "grad_norm": 0.600090503692627,
2990
+ "learning_rate": 6.982407535467416e-08,
2991
+ "loss": 0.0606,
2992
+ "step": 2130
2993
+ },
2994
+ {
2995
+ "epoch": 2.920656634746922,
2996
+ "grad_norm": 0.5937740802764893,
2997
+ "learning_rate": 5.9347155054305104e-08,
2998
+ "loss": 0.0622,
2999
+ "step": 2135
3000
+ },
3001
+ {
3002
+ "epoch": 2.92749658002736,
3003
+ "grad_norm": 0.47966572642326355,
3004
+ "learning_rate": 4.9719862586546374e-08,
3005
+ "loss": 0.0727,
3006
+ "step": 2140
3007
+ },
3008
+ {
3009
+ "epoch": 2.9343365253077973,
3010
+ "grad_norm": 0.5727455615997314,
3011
+ "learning_rate": 4.0942745425179285e-08,
3012
+ "loss": 0.0672,
3013
+ "step": 2145
3014
+ },
3015
+ {
3016
+ "epoch": 2.9411764705882355,
3017
+ "grad_norm": 0.5159140825271606,
3018
+ "learning_rate": 3.3016302697188583e-08,
3019
+ "loss": 0.0669,
3020
+ "step": 2150
3021
+ },
3022
+ {
3023
+ "epoch": 2.948016415868673,
3024
+ "grad_norm": 0.4484996199607849,
3025
+ "learning_rate": 2.5940985154385144e-08,
3026
+ "loss": 0.0488,
3027
+ "step": 2155
3028
+ },
3029
+ {
3030
+ "epoch": 2.954856361149111,
3031
+ "grad_norm": 0.507086455821991,
3032
+ "learning_rate": 1.9717195147776478e-08,
3033
+ "loss": 0.0563,
3034
+ "step": 2160
3035
+ },
3036
+ {
3037
+ "epoch": 2.9616963064295483,
3038
+ "grad_norm": 0.4113939702510834,
3039
+ "learning_rate": 1.4345286604683372e-08,
3040
+ "loss": 0.0604,
3041
+ "step": 2165
3042
+ },
3043
+ {
3044
+ "epoch": 2.9685362517099865,
3045
+ "grad_norm": 0.46114131808280945,
3046
+ "learning_rate": 9.825565008610982e-09,
3047
+ "loss": 0.0665,
3048
+ "step": 2170
3049
+ },
3050
+ {
3051
+ "epoch": 2.975376196990424,
3052
+ "grad_norm": 0.5180654525756836,
3053
+ "learning_rate": 6.158287381879401e-09,
3054
+ "loss": 0.061,
3055
+ "step": 2175
3056
+ },
3057
+ {
3058
+ "epoch": 2.982216142270862,
3059
+ "grad_norm": 0.5392937660217285,
3060
+ "learning_rate": 3.343662271008685e-09,
3061
+ "loss": 0.0649,
3062
+ "step": 2180
3063
+ },
3064
+ {
3065
+ "epoch": 2.9890560875512997,
3066
+ "grad_norm": 0.44279414415359497,
3067
+ "learning_rate": 1.381849734856666e-09,
3068
+ "loss": 0.0482,
3069
+ "step": 2185
3070
+ },
3071
+ {
3072
+ "epoch": 2.9958960328317374,
3073
+ "grad_norm": 0.6033980250358582,
3074
+ "learning_rate": 2.7296133551624193e-10,
3075
+ "loss": 0.0596,
3076
+ "step": 2190
3077
+ },
3078
+ {
3079
+ "epoch": 3.0,
3080
+ "step": 2193,
3081
+ "total_flos": 3.0054141731243295e+18,
3082
+ "train_loss": 0.341107062958062,
3083
+ "train_runtime": 1404.3398,
3084
+ "train_samples_per_second": 49.958,
3085
+ "train_steps_per_second": 1.562
3086
+ }
3087
+ ],
3088
+ "logging_steps": 5,
3089
+ "max_steps": 2193,
3090
+ "num_input_tokens_seen": 0,
3091
+ "num_train_epochs": 3,
3092
+ "save_steps": 20000,
3093
+ "stateful_callbacks": {
3094
+ "TrainerControl": {
3095
+ "args": {
3096
+ "should_epoch_stop": false,
3097
+ "should_evaluate": false,
3098
+ "should_log": false,
3099
+ "should_save": false,
3100
+ "should_training_stop": false
3101
+ },
3102
+ "attributes": {}
3103
+ }
3104
+ },
3105
+ "total_flos": 3.0054141731243295e+18,
3106
+ "train_batch_size": 2,
3107
+ "trial_name": null,
3108
+ "trial_params": null
3109
+ }
instruct/127_128_e3_3e-5/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ccc8d678bd899428b85a0583b6b3c516e9774e43986e9b03dfa9128bf43dde3
3
+ size 8273
instruct/127_128_e3_3e-5/vocab.json ADDED
The diff for this file is too large to render. See raw diff