Hugging Face's logo

RayDu0010
/
tulu10k_klora

Safetensors
Model card Files
xet
 Community
RayDu0010 commited on
Commit
45aab83
·
verified ·
1 Parent(s): f3ef877

Upload folder using huggingface_hub

Browse files
Files changed (14) hide show
  1. instruct/11_128_e3_3e-5/adapter_config.json +39 -0
  2. instruct/11_128_e3_3e-5/adapter_model.safetensors +3 -0
  3. instruct/11_128_e3_3e-5/added_tokens.json +9 -0
  4. instruct/11_128_e3_3e-5/all_results.json +9 -0
  5. instruct/11_128_e3_3e-5/chat_template.jinja +62 -0
  6. instruct/11_128_e3_3e-5/config.json +32 -0
  7. instruct/11_128_e3_3e-5/merges.txt +0 -0
  8. instruct/11_128_e3_3e-5/special_tokens_map.json +33 -0
  9. instruct/11_128_e3_3e-5/tokenizer.json +0 -0
  10. instruct/11_128_e3_3e-5/tokenizer_config.json +234 -0
  11. instruct/11_128_e3_3e-5/train_results.json +9 -0
  12. instruct/11_128_e3_3e-5/trainer_state.json +3046 -0
  13. instruct/11_128_e3_3e-5/training_args.bin +3 -0
  14. instruct/11_128_e3_3e-5/vocab.json +0 -0
instruct/11_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
+ "k_proj",
28
+ "down_proj",
29
+ "q_proj",
30
+ "v_proj",
31
+ "up_proj",
32
+ "gate_proj",
33
+ "o_proj"
34
+ ],
35
+ "task_type": "CAUSAL_LM",
36
+ "trainable_token_indices": null,
37
+ "use_dora": false,
38
+ "use_rslora": false
39
+ }
instruct/11_128_e3_3e-5/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ac74b26622ec3bef81cbf87309888d30dc182127ee2b95720857821bfd77a40
3
+ size 791751704
instruct/11_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/11_128_e3_3e-5/all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 2.6974388809776497e+18,
4
+ "train_loss": 0.37994372332728776,
5
+ "train_runtime": 1294.7645,
6
+ "train_samples": 22903,
7
+ "train_samples_per_second": 53.067,
8
+ "train_steps_per_second": 1.659
9
+ }
instruct/11_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/11_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/11_128_e3_3e-5/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
instruct/11_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/11_128_e3_3e-5/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
instruct/11_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/11_128_e3_3e-5/train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 2.6974388809776497e+18,
4
+ "train_loss": 0.37994372332728776,
5
+ "train_runtime": 1294.7645,
6
+ "train_samples": 22903,
7
+ "train_samples_per_second": 53.067,
8
+ "train_steps_per_second": 1.659
9
+ }
instruct/11_128_e3_3e-5/trainer_state.json ADDED
@@ -0,0 +1,3046 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 2148,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.006983240223463687,
14
+ "grad_norm": 2.461507797241211,
15
+ "learning_rate": 1.111111111111111e-06,
16
+ "loss": 1.5872,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.013966480446927373,
21
+ "grad_norm": 2.4378960132598877,
22
+ "learning_rate": 2.4999999999999998e-06,
23
+ "loss": 1.6627,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.02094972067039106,
28
+ "grad_norm": 1.5489822626113892,
29
+ "learning_rate": 3.888888888888889e-06,
30
+ "loss": 1.5894,
31
+ "step": 15
32
+ },
33
+ {
34
+ "epoch": 0.027932960893854747,
35
+ "grad_norm": 0.9279693961143494,
36
+ "learning_rate": 5.277777777777778e-06,
37
+ "loss": 1.5159,
38
+ "step": 20
39
+ },
40
+ {
41
+ "epoch": 0.034916201117318434,
42
+ "grad_norm": 0.5784654021263123,
43
+ "learning_rate": 6.666666666666667e-06,
44
+ "loss": 1.4935,
45
+ "step": 25
46
+ },
47
+ {
48
+ "epoch": 0.04189944134078212,
49
+ "grad_norm": 0.48185575008392334,
50
+ "learning_rate": 8.055555555555557e-06,
51
+ "loss": 1.3952,
52
+ "step": 30
53
+ },
54
+ {
55
+ "epoch": 0.04888268156424581,
56
+ "grad_norm": 0.430463045835495,
57
+ "learning_rate": 9.444444444444445e-06,
58
+ "loss": 1.4326,
59
+ "step": 35
60
+ },
61
+ {
62
+ "epoch": 0.055865921787709494,
63
+ "grad_norm": 0.39366328716278076,
64
+ "learning_rate": 1.0833333333333334e-05,
65
+ "loss": 1.3799,
66
+ "step": 40
67
+ },
68
+ {
69
+ "epoch": 0.06284916201117319,
70
+ "grad_norm": 0.5764641165733337,
71
+ "learning_rate": 1.2222222222222222e-05,
72
+ "loss": 1.4117,
73
+ "step": 45
74
+ },
75
+ {
76
+ "epoch": 0.06983240223463687,
77
+ "grad_norm": 0.424059122800827,
78
+ "learning_rate": 1.3611111111111111e-05,
79
+ "loss": 1.4345,
80
+ "step": 50
81
+ },
82
+ {
83
+ "epoch": 0.07681564245810056,
84
+ "grad_norm": 0.3716816008090973,
85
+ "learning_rate": 1.5e-05,
86
+ "loss": 1.4249,
87
+ "step": 55
88
+ },
89
+ {
90
+ "epoch": 0.08379888268156424,
91
+ "grad_norm": 0.45050185918807983,
92
+ "learning_rate": 1.638888888888889e-05,
93
+ "loss": 1.3634,
94
+ "step": 60
95
+ },
96
+ {
97
+ "epoch": 0.09078212290502793,
98
+ "grad_norm": 0.38407814502716064,
99
+ "learning_rate": 1.7777777777777777e-05,
100
+ "loss": 1.3728,
101
+ "step": 65
102
+ },
103
+ {
104
+ "epoch": 0.09776536312849161,
105
+ "grad_norm": 0.39405325055122375,
106
+ "learning_rate": 1.9166666666666667e-05,
107
+ "loss": 1.3868,
108
+ "step": 70
109
+ },
110
+ {
111
+ "epoch": 0.10474860335195531,
112
+ "grad_norm": 0.4616144299507141,
113
+ "learning_rate": 2.0555555555555558e-05,
114
+ "loss": 1.3743,
115
+ "step": 75
116
+ },
117
+ {
118
+ "epoch": 0.11173184357541899,
119
+ "grad_norm": 0.4215698540210724,
120
+ "learning_rate": 2.1944444444444445e-05,
121
+ "loss": 1.3424,
122
+ "step": 80
123
+ },
124
+ {
125
+ "epoch": 0.11871508379888268,
126
+ "grad_norm": 0.41426974534988403,
127
+ "learning_rate": 2.3333333333333336e-05,
128
+ "loss": 1.3094,
129
+ "step": 85
130
+ },
131
+ {
132
+ "epoch": 0.12569832402234637,
133
+ "grad_norm": 0.41232773661613464,
134
+ "learning_rate": 2.4722222222222223e-05,
135
+ "loss": 1.2984,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.13268156424581007,
140
+ "grad_norm": 0.4594205617904663,
141
+ "learning_rate": 2.611111111111111e-05,
142
+ "loss": 1.2822,
143
+ "step": 95
144
+ },
145
+ {
146
+ "epoch": 0.13966480446927373,
147
+ "grad_norm": 0.4747633635997772,
148
+ "learning_rate": 2.75e-05,
149
+ "loss": 1.3011,
150
+ "step": 100
151
+ },
152
+ {
153
+ "epoch": 0.14664804469273743,
154
+ "grad_norm": 0.4459817707538605,
155
+ "learning_rate": 2.8888888888888888e-05,
156
+ "loss": 1.2603,
157
+ "step": 105
158
+ },
159
+ {
160
+ "epoch": 0.15363128491620112,
161
+ "grad_norm": 0.47995346784591675,
162
+ "learning_rate": 2.9999982213086704e-05,
163
+ "loss": 1.2383,
164
+ "step": 110
165
+ },
166
+ {
167
+ "epoch": 0.16061452513966482,
168
+ "grad_norm": 0.4630449712276459,
169
+ "learning_rate": 2.9999359675550518e-05,
170
+ "loss": 1.2977,
171
+ "step": 115
172
+ },
173
+ {
174
+ "epoch": 0.16759776536312848,
175
+ "grad_norm": 0.5036991238594055,
176
+ "learning_rate": 2.9997847834532247e-05,
177
+ "loss": 1.2385,
178
+ "step": 120
179
+ },
180
+ {
181
+ "epoch": 0.17458100558659218,
182
+ "grad_norm": 0.5209996104240417,
183
+ "learning_rate": 2.9995446779668078e-05,
184
+ "loss": 1.2,
185
+ "step": 125
186
+ },
187
+ {
188
+ "epoch": 0.18156424581005587,
189
+ "grad_norm": 0.5494832396507263,
190
+ "learning_rate": 2.999215665331518e-05,
191
+ "loss": 1.2306,
192
+ "step": 130
193
+ },
194
+ {
195
+ "epoch": 0.18854748603351956,
196
+ "grad_norm": 0.6769248247146606,
197
+ "learning_rate": 2.9987977650543284e-05,
198
+ "loss": 1.2319,
199
+ "step": 135
200
+ },
201
+ {
202
+ "epoch": 0.19553072625698323,
203
+ "grad_norm": 0.5804729461669922,
204
+ "learning_rate": 2.9982910019123068e-05,
205
+ "loss": 1.2045,
206
+ "step": 140
207
+ },
208
+ {
209
+ "epoch": 0.20251396648044692,
210
+ "grad_norm": 0.6050616502761841,
211
+ "learning_rate": 2.997695405951151e-05,
212
+ "loss": 1.1915,
213
+ "step": 145
214
+ },
215
+ {
216
+ "epoch": 0.20949720670391062,
217
+ "grad_norm": 0.5633885860443115,
218
+ "learning_rate": 2.9970110124834066e-05,
219
+ "loss": 1.1819,
220
+ "step": 150
221
+ },
222
+ {
223
+ "epoch": 0.2164804469273743,
224
+ "grad_norm": 0.6121296882629395,
225
+ "learning_rate": 2.996237862086371e-05,
226
+ "loss": 1.184,
227
+ "step": 155
228
+ },
229
+ {
230
+ "epoch": 0.22346368715083798,
231
+ "grad_norm": 0.5915355682373047,
232
+ "learning_rate": 2.995376000599692e-05,
233
+ "loss": 1.0977,
234
+ "step": 160
235
+ },
236
+ {
237
+ "epoch": 0.23044692737430167,
238
+ "grad_norm": 2.789428949356079,
239
+ "learning_rate": 2.9944254791226444e-05,
240
+ "loss": 1.1485,
241
+ "step": 165
242
+ },
243
+ {
244
+ "epoch": 0.23743016759776536,
245
+ "grad_norm": 2.3790037631988525,
246
+ "learning_rate": 2.9933863540111047e-05,
247
+ "loss": 1.0759,
248
+ "step": 170
249
+ },
250
+ {
251
+ "epoch": 0.24441340782122906,
252
+ "grad_norm": 0.6987988352775574,
253
+ "learning_rate": 2.9922586868742085e-05,
254
+ "loss": 1.0949,
255
+ "step": 175
256
+ },
257
+ {
258
+ "epoch": 0.25139664804469275,
259
+ "grad_norm": 0.6527789831161499,
260
+ "learning_rate": 2.991042544570697e-05,
261
+ "loss": 1.1081,
262
+ "step": 180
263
+ },
264
+ {
265
+ "epoch": 0.25837988826815644,
266
+ "grad_norm": 0.6399208307266235,
267
+ "learning_rate": 2.9897379992049543e-05,
268
+ "loss": 1.0519,
269
+ "step": 185
270
+ },
271
+ {
272
+ "epoch": 0.26536312849162014,
273
+ "grad_norm": 0.7913739085197449,
274
+ "learning_rate": 2.988345128122731e-05,
275
+ "loss": 1.0851,
276
+ "step": 190
277
+ },
278
+ {
279
+ "epoch": 0.2723463687150838,
280
+ "grad_norm": 0.7076494097709656,
281
+ "learning_rate": 2.9868640139065595e-05,
282
+ "loss": 1.0573,
283
+ "step": 195
284
+ },
285
+ {
286
+ "epoch": 0.27932960893854747,
287
+ "grad_norm": 0.8066948652267456,
288
+ "learning_rate": 2.985294744370857e-05,
289
+ "loss": 1.0933,
290
+ "step": 200
291
+ },
292
+ {
293
+ "epoch": 0.28631284916201116,
294
+ "grad_norm": 0.7052973508834839,
295
+ "learning_rate": 2.9836374125567193e-05,
296
+ "loss": 1.0261,
297
+ "step": 205
298
+ },
299
+ {
300
+ "epoch": 0.29329608938547486,
301
+ "grad_norm": 0.7760833501815796,
302
+ "learning_rate": 2.981892116726404e-05,
303
+ "loss": 1.0084,
304
+ "step": 210
305
+ },
306
+ {
307
+ "epoch": 0.30027932960893855,
308
+ "grad_norm": 0.7197896838188171,
309
+ "learning_rate": 2.9800589603575072e-05,
310
+ "loss": 1.0341,
311
+ "step": 215
312
+ },
313
+ {
314
+ "epoch": 0.30726256983240224,
315
+ "grad_norm": 0.7907305359840393,
316
+ "learning_rate": 2.978138052136824e-05,
317
+ "loss": 1.0414,
318
+ "step": 220
319
+ },
320
+ {
321
+ "epoch": 0.31424581005586594,
322
+ "grad_norm": 0.8101890087127686,
323
+ "learning_rate": 2.976129505953907e-05,
324
+ "loss": 1.0368,
325
+ "step": 225
326
+ },
327
+ {
328
+ "epoch": 0.32122905027932963,
329
+ "grad_norm": 0.8439146280288696,
330
+ "learning_rate": 2.974033440894314e-05,
331
+ "loss": 1.0059,
332
+ "step": 230
333
+ },
334
+ {
335
+ "epoch": 0.32821229050279327,
336
+ "grad_norm": 0.8499584794044495,
337
+ "learning_rate": 2.9718499812325478e-05,
338
+ "loss": 0.9483,
339
+ "step": 235
340
+ },
341
+ {
342
+ "epoch": 0.33519553072625696,
343
+ "grad_norm": 0.8462314605712891,
344
+ "learning_rate": 2.9695792564246864e-05,
345
+ "loss": 1.0279,
346
+ "step": 240
347
+ },
348
+ {
349
+ "epoch": 0.34217877094972066,
350
+ "grad_norm": 0.8033967614173889,
351
+ "learning_rate": 2.9672214011007087e-05,
352
+ "loss": 0.9399,
353
+ "step": 245
354
+ },
355
+ {
356
+ "epoch": 0.34916201117318435,
357
+ "grad_norm": 0.9281762838363647,
358
+ "learning_rate": 2.964776555056512e-05,
359
+ "loss": 0.9564,
360
+ "step": 250
361
+ },
362
+ {
363
+ "epoch": 0.35614525139664804,
364
+ "grad_norm": 0.9242972731590271,
365
+ "learning_rate": 2.9622448632456253e-05,
366
+ "loss": 0.9497,
367
+ "step": 255
368
+ },
369
+ {
370
+ "epoch": 0.36312849162011174,
371
+ "grad_norm": 0.900638222694397,
372
+ "learning_rate": 2.9596264757706114e-05,
373
+ "loss": 0.9841,
374
+ "step": 260
375
+ },
376
+ {
377
+ "epoch": 0.37011173184357543,
378
+ "grad_norm": 1.006266474723816,
379
+ "learning_rate": 2.9569215478741706e-05,
380
+ "loss": 0.8664,
381
+ "step": 265
382
+ },
383
+ {
384
+ "epoch": 0.3770949720670391,
385
+ "grad_norm": 0.8874937891960144,
386
+ "learning_rate": 2.954130239929936e-05,
387
+ "loss": 0.9254,
388
+ "step": 270
389
+ },
390
+ {
391
+ "epoch": 0.3840782122905028,
392
+ "grad_norm": 0.9324424266815186,
393
+ "learning_rate": 2.9512527174329636e-05,
394
+ "loss": 0.8327,
395
+ "step": 275
396
+ },
397
+ {
398
+ "epoch": 0.39106145251396646,
399
+ "grad_norm": 1.0240839719772339,
400
+ "learning_rate": 2.9482891509899222e-05,
401
+ "loss": 0.8354,
402
+ "step": 280
403
+ },
404
+ {
405
+ "epoch": 0.39804469273743015,
406
+ "grad_norm": 0.8711543679237366,
407
+ "learning_rate": 2.9452397163089748e-05,
408
+ "loss": 0.9261,
409
+ "step": 285
410
+ },
411
+ {
412
+ "epoch": 0.40502793296089384,
413
+ "grad_norm": 0.9490526914596558,
414
+ "learning_rate": 2.9421045941893656e-05,
415
+ "loss": 0.8336,
416
+ "step": 290
417
+ },
418
+ {
419
+ "epoch": 0.41201117318435754,
420
+ "grad_norm": 0.9295695424079895,
421
+ "learning_rate": 2.9388839705106975e-05,
422
+ "loss": 0.8996,
423
+ "step": 295
424
+ },
425
+ {
426
+ "epoch": 0.41899441340782123,
427
+ "grad_norm": 0.9833492636680603,
428
+ "learning_rate": 2.9355780362219115e-05,
429
+ "loss": 0.8445,
430
+ "step": 300
431
+ },
432
+ {
433
+ "epoch": 0.4259776536312849,
434
+ "grad_norm": 0.9425599575042725,
435
+ "learning_rate": 2.9321869873299666e-05,
436
+ "loss": 0.8421,
437
+ "step": 305
438
+ },
439
+ {
440
+ "epoch": 0.4329608938547486,
441
+ "grad_norm": 1.0283668041229248,
442
+ "learning_rate": 2.928711024888218e-05,
443
+ "loss": 0.8502,
444
+ "step": 310
445
+ },
446
+ {
447
+ "epoch": 0.4399441340782123,
448
+ "grad_norm": 1.1260733604431152,
449
+ "learning_rate": 2.9251503549844964e-05,
450
+ "loss": 0.8282,
451
+ "step": 315
452
+ },
453
+ {
454
+ "epoch": 0.44692737430167595,
455
+ "grad_norm": 1.0821149349212646,
456
+ "learning_rate": 2.921505188728891e-05,
457
+ "loss": 0.8322,
458
+ "step": 320
459
+ },
460
+ {
461
+ "epoch": 0.45391061452513964,
462
+ "grad_norm": 0.8807809352874756,
463
+ "learning_rate": 2.9177757422412294e-05,
464
+ "loss": 0.8347,
465
+ "step": 325
466
+ },
467
+ {
468
+ "epoch": 0.46089385474860334,
469
+ "grad_norm": 1.1025316715240479,
470
+ "learning_rate": 2.913962236638268e-05,
471
+ "loss": 0.8039,
472
+ "step": 330
473
+ },
474
+ {
475
+ "epoch": 0.46787709497206703,
476
+ "grad_norm": 1.0225765705108643,
477
+ "learning_rate": 2.9100648980205787e-05,
478
+ "loss": 0.8399,
479
+ "step": 335
480
+ },
481
+ {
482
+ "epoch": 0.4748603351955307,
483
+ "grad_norm": 0.9346492886543274,
484
+ "learning_rate": 2.906083957459146e-05,
485
+ "loss": 0.7627,
486
+ "step": 340
487
+ },
488
+ {
489
+ "epoch": 0.4818435754189944,
490
+ "grad_norm": 1.1114012002944946,
491
+ "learning_rate": 2.9020196509816662e-05,
492
+ "loss": 0.829,
493
+ "step": 345
494
+ },
495
+ {
496
+ "epoch": 0.4888268156424581,
497
+ "grad_norm": 1.0867840051651,
498
+ "learning_rate": 2.897872219558552e-05,
499
+ "loss": 0.7858,
500
+ "step": 350
501
+ },
502
+ {
503
+ "epoch": 0.4958100558659218,
504
+ "grad_norm": 1.0677963495254517,
505
+ "learning_rate": 2.8936419090886484e-05,
506
+ "loss": 0.7856,
507
+ "step": 355
508
+ },
509
+ {
510
+ "epoch": 0.5027932960893855,
511
+ "grad_norm": 1.162087082862854,
512
+ "learning_rate": 2.8893289703846503e-05,
513
+ "loss": 0.7535,
514
+ "step": 360
515
+ },
516
+ {
517
+ "epoch": 0.5097765363128491,
518
+ "grad_norm": 1.0000637769699097,
519
+ "learning_rate": 2.884933659158234e-05,
520
+ "loss": 0.7684,
521
+ "step": 365
522
+ },
523
+ {
524
+ "epoch": 0.5167597765363129,
525
+ "grad_norm": 1.057491660118103,
526
+ "learning_rate": 2.8804562360048964e-05,
527
+ "loss": 0.809,
528
+ "step": 370
529
+ },
530
+ {
531
+ "epoch": 0.5237430167597765,
532
+ "grad_norm": 1.1435449123382568,
533
+ "learning_rate": 2.8758969663885026e-05,
534
+ "loss": 0.763,
535
+ "step": 375
536
+ },
537
+ {
538
+ "epoch": 0.5307262569832403,
539
+ "grad_norm": 1.0342165231704712,
540
+ "learning_rate": 2.8712561206255484e-05,
541
+ "loss": 0.7581,
542
+ "step": 380
543
+ },
544
+ {
545
+ "epoch": 0.5377094972067039,
546
+ "grad_norm": 1.1961387395858765,
547
+ "learning_rate": 2.8665339738691326e-05,
548
+ "loss": 0.769,
549
+ "step": 385
550
+ },
551
+ {
552
+ "epoch": 0.5446927374301676,
553
+ "grad_norm": 1.283997893333435,
554
+ "learning_rate": 2.8617308060926443e-05,
555
+ "loss": 0.711,
556
+ "step": 390
557
+ },
558
+ {
559
+ "epoch": 0.5516759776536313,
560
+ "grad_norm": 1.0753425359725952,
561
+ "learning_rate": 2.856846902073161e-05,
562
+ "loss": 0.758,
563
+ "step": 395
564
+ },
565
+ {
566
+ "epoch": 0.5586592178770949,
567
+ "grad_norm": 1.044743537902832,
568
+ "learning_rate": 2.851882551374568e-05,
569
+ "loss": 0.7187,
570
+ "step": 400
571
+ },
572
+ {
573
+ "epoch": 0.5656424581005587,
574
+ "grad_norm": 1.2359068393707275,
575
+ "learning_rate": 2.8468380483303873e-05,
576
+ "loss": 0.697,
577
+ "step": 405
578
+ },
579
+ {
580
+ "epoch": 0.5726256983240223,
581
+ "grad_norm": 1.088114857673645,
582
+ "learning_rate": 2.8417136920263277e-05,
583
+ "loss": 0.7264,
584
+ "step": 410
585
+ },
586
+ {
587
+ "epoch": 0.5796089385474861,
588
+ "grad_norm": 1.1865432262420654,
589
+ "learning_rate": 2.8365097862825516e-05,
590
+ "loss": 0.6896,
591
+ "step": 415
592
+ },
593
+ {
594
+ "epoch": 0.5865921787709497,
595
+ "grad_norm": 1.1307533979415894,
596
+ "learning_rate": 2.8312266396356647e-05,
597
+ "loss": 0.6702,
598
+ "step": 420
599
+ },
600
+ {
601
+ "epoch": 0.5935754189944135,
602
+ "grad_norm": 1.1586530208587646,
603
+ "learning_rate": 2.825864565320418e-05,
604
+ "loss": 0.669,
605
+ "step": 425
606
+ },
607
+ {
608
+ "epoch": 0.6005586592178771,
609
+ "grad_norm": 1.1601227521896362,
610
+ "learning_rate": 2.8204238812511405e-05,
611
+ "loss": 0.7077,
612
+ "step": 430
613
+ },
614
+ {
615
+ "epoch": 0.6075418994413407,
616
+ "grad_norm": 1.266157627105713,
617
+ "learning_rate": 2.814904910002888e-05,
618
+ "loss": 0.6346,
619
+ "step": 435
620
+ },
621
+ {
622
+ "epoch": 0.6145251396648045,
623
+ "grad_norm": 1.1432222127914429,
624
+ "learning_rate": 2.8093079787923186e-05,
625
+ "loss": 0.6692,
626
+ "step": 440
627
+ },
628
+ {
629
+ "epoch": 0.6215083798882681,
630
+ "grad_norm": 0.9997245073318481,
631
+ "learning_rate": 2.8036334194582924e-05,
632
+ "loss": 0.6771,
633
+ "step": 445
634
+ },
635
+ {
636
+ "epoch": 0.6284916201117319,
637
+ "grad_norm": 1.1972033977508545,
638
+ "learning_rate": 2.797881568442197e-05,
639
+ "loss": 0.6692,
640
+ "step": 450
641
+ },
642
+ {
643
+ "epoch": 0.6354748603351955,
644
+ "grad_norm": 1.2365937232971191,
645
+ "learning_rate": 2.7920527667679984e-05,
646
+ "loss": 0.6527,
647
+ "step": 455
648
+ },
649
+ {
650
+ "epoch": 0.6424581005586593,
651
+ "grad_norm": 1.1503313779830933,
652
+ "learning_rate": 2.7861473600220245e-05,
653
+ "loss": 0.6581,
654
+ "step": 460
655
+ },
656
+ {
657
+ "epoch": 0.6494413407821229,
658
+ "grad_norm": 1.263296127319336,
659
+ "learning_rate": 2.7801656983324738e-05,
660
+ "loss": 0.6549,
661
+ "step": 465
662
+ },
663
+ {
664
+ "epoch": 0.6564245810055865,
665
+ "grad_norm": 1.081668496131897,
666
+ "learning_rate": 2.7741081363486574e-05,
667
+ "loss": 0.6637,
668
+ "step": 470
669
+ },
670
+ {
671
+ "epoch": 0.6634078212290503,
672
+ "grad_norm": 1.1281845569610596,
673
+ "learning_rate": 2.7679750332199707e-05,
674
+ "loss": 0.616,
675
+ "step": 475
676
+ },
677
+ {
678
+ "epoch": 0.6703910614525139,
679
+ "grad_norm": 1.0776785612106323,
680
+ "learning_rate": 2.7617667525746006e-05,
681
+ "loss": 0.6386,
682
+ "step": 480
683
+ },
684
+ {
685
+ "epoch": 0.6773743016759777,
686
+ "grad_norm": 1.2409050464630127,
687
+ "learning_rate": 2.7554836624979666e-05,
688
+ "loss": 0.572,
689
+ "step": 485
690
+ },
691
+ {
692
+ "epoch": 0.6843575418994413,
693
+ "grad_norm": 1.0798362493515015,
694
+ "learning_rate": 2.7491261355108974e-05,
695
+ "loss": 0.5536,
696
+ "step": 490
697
+ },
698
+ {
699
+ "epoch": 0.6913407821229051,
700
+ "grad_norm": 1.2088700532913208,
701
+ "learning_rate": 2.7426945485475416e-05,
702
+ "loss": 0.5962,
703
+ "step": 495
704
+ },
705
+ {
706
+ "epoch": 0.6983240223463687,
707
+ "grad_norm": 1.0456690788269043,
708
+ "learning_rate": 2.7361892829330234e-05,
709
+ "loss": 0.6053,
710
+ "step": 500
711
+ },
712
+ {
713
+ "epoch": 0.7053072625698324,
714
+ "grad_norm": 1.0685166120529175,
715
+ "learning_rate": 2.729610724360832e-05,
716
+ "loss": 0.6051,
717
+ "step": 505
718
+ },
719
+ {
720
+ "epoch": 0.7122905027932961,
721
+ "grad_norm": 1.2368314266204834,
722
+ "learning_rate": 2.7229592628699528e-05,
723
+ "loss": 0.6327,
724
+ "step": 510
725
+ },
726
+ {
727
+ "epoch": 0.7192737430167597,
728
+ "grad_norm": 1.2205802202224731,
729
+ "learning_rate": 2.716235292821745e-05,
730
+ "loss": 0.6162,
731
+ "step": 515
732
+ },
733
+ {
734
+ "epoch": 0.7262569832402235,
735
+ "grad_norm": 1.286219596862793,
736
+ "learning_rate": 2.7094392128765597e-05,
737
+ "loss": 0.6092,
738
+ "step": 520
739
+ },
740
+ {
741
+ "epoch": 0.7332402234636871,
742
+ "grad_norm": 1.4330753087997437,
743
+ "learning_rate": 2.7025714259701e-05,
744
+ "loss": 0.5622,
745
+ "step": 525
746
+ },
747
+ {
748
+ "epoch": 0.7402234636871509,
749
+ "grad_norm": 1.1543325185775757,
750
+ "learning_rate": 2.6956323392895366e-05,
751
+ "loss": 0.5963,
752
+ "step": 530
753
+ },
754
+ {
755
+ "epoch": 0.7472067039106145,
756
+ "grad_norm": 1.1962709426879883,
757
+ "learning_rate": 2.688622364249361e-05,
758
+ "loss": 0.5991,
759
+ "step": 535
760
+ },
761
+ {
762
+ "epoch": 0.7541899441340782,
763
+ "grad_norm": 1.2152200937271118,
764
+ "learning_rate": 2.6815419164669964e-05,
765
+ "loss": 0.5928,
766
+ "step": 540
767
+ },
768
+ {
769
+ "epoch": 0.7611731843575419,
770
+ "grad_norm": 1.3667744398117065,
771
+ "learning_rate": 2.6743914157381542e-05,
772
+ "loss": 0.5689,
773
+ "step": 545
774
+ },
775
+ {
776
+ "epoch": 0.7681564245810056,
777
+ "grad_norm": 1.2070188522338867,
778
+ "learning_rate": 2.6671712860119454e-05,
779
+ "loss": 0.5498,
780
+ "step": 550
781
+ },
782
+ {
783
+ "epoch": 0.7751396648044693,
784
+ "grad_norm": 1.1513986587524414,
785
+ "learning_rate": 2.6598819553657447e-05,
786
+ "loss": 0.5705,
787
+ "step": 555
788
+ },
789
+ {
790
+ "epoch": 0.7821229050279329,
791
+ "grad_norm": 1.138221025466919,
792
+ "learning_rate": 2.6525238559798087e-05,
793
+ "loss": 0.4956,
794
+ "step": 560
795
+ },
796
+ {
797
+ "epoch": 0.7891061452513967,
798
+ "grad_norm": 1.2426525354385376,
799
+ "learning_rate": 2.6450974241116545e-05,
800
+ "loss": 0.4789,
801
+ "step": 565
802
+ },
803
+ {
804
+ "epoch": 0.7960893854748603,
805
+ "grad_norm": 1.1912192106246948,
806
+ "learning_rate": 2.6376031000701934e-05,
807
+ "loss": 0.5078,
808
+ "step": 570
809
+ },
810
+ {
811
+ "epoch": 0.803072625698324,
812
+ "grad_norm": 1.195440649986267,
813
+ "learning_rate": 2.630041328189623e-05,
814
+ "loss": 0.5732,
815
+ "step": 575
816
+ },
817
+ {
818
+ "epoch": 0.8100558659217877,
819
+ "grad_norm": 1.0765756368637085,
820
+ "learning_rate": 2.6224125568030872e-05,
821
+ "loss": 0.5452,
822
+ "step": 580
823
+ },
824
+ {
825
+ "epoch": 0.8170391061452514,
826
+ "grad_norm": 1.2104696035385132,
827
+ "learning_rate": 2.6147172382160913e-05,
828
+ "loss": 0.4366,
829
+ "step": 585
830
+ },
831
+ {
832
+ "epoch": 0.8240223463687151,
833
+ "grad_norm": 1.1734176874160767,
834
+ "learning_rate": 2.606955828679686e-05,
835
+ "loss": 0.5445,
836
+ "step": 590
837
+ },
838
+ {
839
+ "epoch": 0.8310055865921788,
840
+ "grad_norm": 1.389171838760376,
841
+ "learning_rate": 2.5991287883634173e-05,
842
+ "loss": 0.4752,
843
+ "step": 595
844
+ },
845
+ {
846
+ "epoch": 0.8379888268156425,
847
+ "grad_norm": 1.2040082216262817,
848
+ "learning_rate": 2.5912365813280416e-05,
849
+ "loss": 0.4541,
850
+ "step": 600
851
+ },
852
+ {
853
+ "epoch": 0.8449720670391061,
854
+ "grad_norm": 1.1748125553131104,
855
+ "learning_rate": 2.5832796754980138e-05,
856
+ "loss": 0.5238,
857
+ "step": 605
858
+ },
859
+ {
860
+ "epoch": 0.8519553072625698,
861
+ "grad_norm": 1.2526957988739014,
862
+ "learning_rate": 2.5752585426337443e-05,
863
+ "loss": 0.4905,
864
+ "step": 610
865
+ },
866
+ {
867
+ "epoch": 0.8589385474860335,
868
+ "grad_norm": 1.1745952367782593,
869
+ "learning_rate": 2.5671736583036256e-05,
870
+ "loss": 0.4869,
871
+ "step": 615
872
+ },
873
+ {
874
+ "epoch": 0.8659217877094972,
875
+ "grad_norm": 1.1998168230056763,
876
+ "learning_rate": 2.55902550185584e-05,
877
+ "loss": 0.4778,
878
+ "step": 620
879
+ },
880
+ {
881
+ "epoch": 0.8729050279329609,
882
+ "grad_norm": 1.2667831182479858,
883
+ "learning_rate": 2.5508145563899364e-05,
884
+ "loss": 0.4669,
885
+ "step": 625
886
+ },
887
+ {
888
+ "epoch": 0.8798882681564246,
889
+ "grad_norm": 1.152965784072876,
890
+ "learning_rate": 2.5425413087281898e-05,
891
+ "loss": 0.4833,
892
+ "step": 630
893
+ },
894
+ {
895
+ "epoch": 0.8868715083798883,
896
+ "grad_norm": 1.1805378198623657,
897
+ "learning_rate": 2.534206249386737e-05,
898
+ "loss": 0.4633,
899
+ "step": 635
900
+ },
901
+ {
902
+ "epoch": 0.8938547486033519,
903
+ "grad_norm": 1.3154098987579346,
904
+ "learning_rate": 2.525809872546492e-05,
905
+ "loss": 0.461,
906
+ "step": 640
907
+ },
908
+ {
909
+ "epoch": 0.9008379888268156,
910
+ "grad_norm": 1.2145469188690186,
911
+ "learning_rate": 2.517352676023851e-05,
912
+ "loss": 0.4741,
913
+ "step": 645
914
+ },
915
+ {
916
+ "epoch": 0.9078212290502793,
917
+ "grad_norm": 1.1841624975204468,
918
+ "learning_rate": 2.5088351612411727e-05,
919
+ "loss": 0.4574,
920
+ "step": 650
921
+ },
922
+ {
923
+ "epoch": 0.914804469273743,
924
+ "grad_norm": 1.334181308746338,
925
+ "learning_rate": 2.500257833197052e-05,
926
+ "loss": 0.4614,
927
+ "step": 655
928
+ },
929
+ {
930
+ "epoch": 0.9217877094972067,
931
+ "grad_norm": 1.1441943645477295,
932
+ "learning_rate": 2.4916212004363775e-05,
933
+ "loss": 0.4279,
934
+ "step": 660
935
+ },
936
+ {
937
+ "epoch": 0.9287709497206704,
938
+ "grad_norm": 1.1630587577819824,
939
+ "learning_rate": 2.4829257750201806e-05,
940
+ "loss": 0.4757,
941
+ "step": 665
942
+ },
943
+ {
944
+ "epoch": 0.9357541899441341,
945
+ "grad_norm": 1.3652783632278442,
946
+ "learning_rate": 2.4741720724952754e-05,
947
+ "loss": 0.4777,
948
+ "step": 670
949
+ },
950
+ {
951
+ "epoch": 0.9427374301675978,
952
+ "grad_norm": 1.1949814558029175,
953
+ "learning_rate": 2.465360611863694e-05,
954
+ "loss": 0.4359,
955
+ "step": 675
956
+ },
957
+ {
958
+ "epoch": 0.9497206703910615,
959
+ "grad_norm": 1.205856442451477,
960
+ "learning_rate": 2.4564919155519116e-05,
961
+ "loss": 0.391,
962
+ "step": 680
963
+ },
964
+ {
965
+ "epoch": 0.9567039106145251,
966
+ "grad_norm": 1.141206979751587,
967
+ "learning_rate": 2.4475665093798766e-05,
968
+ "loss": 0.4373,
969
+ "step": 685
970
+ },
971
+ {
972
+ "epoch": 0.9636871508379888,
973
+ "grad_norm": 1.270096778869629,
974
+ "learning_rate": 2.4385849225298303e-05,
975
+ "loss": 0.4261,
976
+ "step": 690
977
+ },
978
+ {
979
+ "epoch": 0.9706703910614525,
980
+ "grad_norm": 1.30154550075531,
981
+ "learning_rate": 2.4295476875149358e-05,
982
+ "loss": 0.413,
983
+ "step": 695
984
+ },
985
+ {
986
+ "epoch": 0.9776536312849162,
987
+ "grad_norm": 1.2080790996551514,
988
+ "learning_rate": 2.420455340147704e-05,
989
+ "loss": 0.4982,
990
+ "step": 700
991
+ },
992
+ {
993
+ "epoch": 0.9846368715083799,
994
+ "grad_norm": 1.2074291706085205,
995
+ "learning_rate": 2.4113084195082256e-05,
996
+ "loss": 0.4947,
997
+ "step": 705
998
+ },
999
+ {
1000
+ "epoch": 0.9916201117318436,
1001
+ "grad_norm": 1.2935150861740112,
1002
+ "learning_rate": 2.402107467912209e-05,
1003
+ "loss": 0.4183,
1004
+ "step": 710
1005
+ },
1006
+ {
1007
+ "epoch": 0.9986033519553073,
1008
+ "grad_norm": 1.1344910860061646,
1009
+ "learning_rate": 2.3928530308788282e-05,
1010
+ "loss": 0.4575,
1011
+ "step": 715
1012
+ },
1013
+ {
1014
+ "epoch": 1.005586592178771,
1015
+ "grad_norm": 1.0799304246902466,
1016
+ "learning_rate": 2.3835456570983777e-05,
1017
+ "loss": 0.3445,
1018
+ "step": 720
1019
+ },
1020
+ {
1021
+ "epoch": 1.0125698324022347,
1022
+ "grad_norm": 1.3482083082199097,
1023
+ "learning_rate": 2.3741858983997415e-05,
1024
+ "loss": 0.3868,
1025
+ "step": 725
1026
+ },
1027
+ {
1028
+ "epoch": 1.0195530726256983,
1029
+ "grad_norm": 1.1869895458221436,
1030
+ "learning_rate": 2.3647743097176758e-05,
1031
+ "loss": 0.324,
1032
+ "step": 730
1033
+ },
1034
+ {
1035
+ "epoch": 1.026536312849162,
1036
+ "grad_norm": 1.2626792192459106,
1037
+ "learning_rate": 2.355311449059906e-05,
1038
+ "loss": 0.3511,
1039
+ "step": 735
1040
+ },
1041
+ {
1042
+ "epoch": 1.0335195530726258,
1043
+ "grad_norm": 1.2026513814926147,
1044
+ "learning_rate": 2.3457978774740455e-05,
1045
+ "loss": 0.3552,
1046
+ "step": 740
1047
+ },
1048
+ {
1049
+ "epoch": 1.0405027932960893,
1050
+ "grad_norm": 1.2248936891555786,
1051
+ "learning_rate": 2.3362341590143284e-05,
1052
+ "loss": 0.3612,
1053
+ "step": 745
1054
+ },
1055
+ {
1056
+ "epoch": 1.047486033519553,
1057
+ "grad_norm": 1.1500792503356934,
1058
+ "learning_rate": 2.3266208607081686e-05,
1059
+ "loss": 0.3113,
1060
+ "step": 750
1061
+ },
1062
+ {
1063
+ "epoch": 1.0544692737430168,
1064
+ "grad_norm": 1.2256200313568115,
1065
+ "learning_rate": 2.316958552522541e-05,
1066
+ "loss": 0.3235,
1067
+ "step": 755
1068
+ },
1069
+ {
1070
+ "epoch": 1.0614525139664805,
1071
+ "grad_norm": 1.3231208324432373,
1072
+ "learning_rate": 2.307247807330188e-05,
1073
+ "loss": 0.3049,
1074
+ "step": 760
1075
+ },
1076
+ {
1077
+ "epoch": 1.068435754189944,
1078
+ "grad_norm": 1.1748385429382324,
1079
+ "learning_rate": 2.297489200875654e-05,
1080
+ "loss": 0.3256,
1081
+ "step": 765
1082
+ },
1083
+ {
1084
+ "epoch": 1.0754189944134078,
1085
+ "grad_norm": 1.2670568227767944,
1086
+ "learning_rate": 2.2876833117411527e-05,
1087
+ "loss": 0.3514,
1088
+ "step": 770
1089
+ },
1090
+ {
1091
+ "epoch": 1.0824022346368716,
1092
+ "grad_norm": 1.1951897144317627,
1093
+ "learning_rate": 2.277830721312258e-05,
1094
+ "loss": 0.3371,
1095
+ "step": 775
1096
+ },
1097
+ {
1098
+ "epoch": 1.089385474860335,
1099
+ "grad_norm": 2.1150286197662354,
1100
+ "learning_rate": 2.267932013743439e-05,
1101
+ "loss": 0.3015,
1102
+ "step": 780
1103
+ },
1104
+ {
1105
+ "epoch": 1.0963687150837989,
1106
+ "grad_norm": 1.2811150550842285,
1107
+ "learning_rate": 2.2579877759234222e-05,
1108
+ "loss": 0.3588,
1109
+ "step": 785
1110
+ },
1111
+ {
1112
+ "epoch": 1.1033519553072626,
1113
+ "grad_norm": 1.282172679901123,
1114
+ "learning_rate": 2.247998597440398e-05,
1115
+ "loss": 0.321,
1116
+ "step": 790
1117
+ },
1118
+ {
1119
+ "epoch": 1.1103351955307263,
1120
+ "grad_norm": 1.0987186431884766,
1121
+ "learning_rate": 2.237965070547062e-05,
1122
+ "loss": 0.3497,
1123
+ "step": 795
1124
+ },
1125
+ {
1126
+ "epoch": 1.1173184357541899,
1127
+ "grad_norm": 1.262815237045288,
1128
+ "learning_rate": 2.2278877901255028e-05,
1129
+ "loss": 0.3124,
1130
+ "step": 800
1131
+ },
1132
+ {
1133
+ "epoch": 1.1243016759776536,
1134
+ "grad_norm": 1.2610952854156494,
1135
+ "learning_rate": 2.2177673536519297e-05,
1136
+ "loss": 0.3107,
1137
+ "step": 805
1138
+ },
1139
+ {
1140
+ "epoch": 1.1312849162011174,
1141
+ "grad_norm": 1.1604728698730469,
1142
+ "learning_rate": 2.2076043611612507e-05,
1143
+ "loss": 0.3046,
1144
+ "step": 810
1145
+ },
1146
+ {
1147
+ "epoch": 1.138268156424581,
1148
+ "grad_norm": 1.3400182723999023,
1149
+ "learning_rate": 2.1973994152114954e-05,
1150
+ "loss": 0.3078,
1151
+ "step": 815
1152
+ },
1153
+ {
1154
+ "epoch": 1.1452513966480447,
1155
+ "grad_norm": 1.1124814748764038,
1156
+ "learning_rate": 2.18715312084809e-05,
1157
+ "loss": 0.2894,
1158
+ "step": 820
1159
+ },
1160
+ {
1161
+ "epoch": 1.1522346368715084,
1162
+ "grad_norm": 1.239890694618225,
1163
+ "learning_rate": 2.1768660855679844e-05,
1164
+ "loss": 0.2935,
1165
+ "step": 825
1166
+ },
1167
+ {
1168
+ "epoch": 1.1592178770949721,
1169
+ "grad_norm": 1.380743384361267,
1170
+ "learning_rate": 2.166538919283634e-05,
1171
+ "loss": 0.3135,
1172
+ "step": 830
1173
+ },
1174
+ {
1175
+ "epoch": 1.1662011173184357,
1176
+ "grad_norm": 1.1818221807479858,
1177
+ "learning_rate": 2.1561722342868397e-05,
1178
+ "loss": 0.2969,
1179
+ "step": 835
1180
+ },
1181
+ {
1182
+ "epoch": 1.1731843575418994,
1183
+ "grad_norm": 1.0518245697021484,
1184
+ "learning_rate": 2.145766645212443e-05,
1185
+ "loss": 0.2955,
1186
+ "step": 840
1187
+ },
1188
+ {
1189
+ "epoch": 1.1801675977653632,
1190
+ "grad_norm": 1.1569312810897827,
1191
+ "learning_rate": 2.1353227690018865e-05,
1192
+ "loss": 0.3056,
1193
+ "step": 845
1194
+ },
1195
+ {
1196
+ "epoch": 1.1871508379888267,
1197
+ "grad_norm": 1.1667453050613403,
1198
+ "learning_rate": 2.124841224866636e-05,
1199
+ "loss": 0.3197,
1200
+ "step": 850
1201
+ },
1202
+ {
1203
+ "epoch": 1.1941340782122905,
1204
+ "grad_norm": 1.1288772821426392,
1205
+ "learning_rate": 2.1143226342514658e-05,
1206
+ "loss": 0.2527,
1207
+ "step": 855
1208
+ },
1209
+ {
1210
+ "epoch": 1.2011173184357542,
1211
+ "grad_norm": 1.1621415615081787,
1212
+ "learning_rate": 2.103767620797615e-05,
1213
+ "loss": 0.2998,
1214
+ "step": 860
1215
+ },
1216
+ {
1217
+ "epoch": 1.208100558659218,
1218
+ "grad_norm": 1.292885184288025,
1219
+ "learning_rate": 2.0931768103058137e-05,
1220
+ "loss": 0.2801,
1221
+ "step": 865
1222
+ },
1223
+ {
1224
+ "epoch": 1.2150837988826815,
1225
+ "grad_norm": 1.3309292793273926,
1226
+ "learning_rate": 2.082550830699176e-05,
1227
+ "loss": 0.2755,
1228
+ "step": 870
1229
+ },
1230
+ {
1231
+ "epoch": 1.2220670391061452,
1232
+ "grad_norm": 1.195035457611084,
1233
+ "learning_rate": 2.0718903119859743e-05,
1234
+ "loss": 0.274,
1235
+ "step": 875
1236
+ },
1237
+ {
1238
+ "epoch": 1.229050279329609,
1239
+ "grad_norm": 1.2508939504623413,
1240
+ "learning_rate": 2.061195886222283e-05,
1241
+ "loss": 0.2773,
1242
+ "step": 880
1243
+ },
1244
+ {
1245
+ "epoch": 1.2360335195530725,
1246
+ "grad_norm": 1.5574023723602295,
1247
+ "learning_rate": 2.0504681874745082e-05,
1248
+ "loss": 0.2559,
1249
+ "step": 885
1250
+ },
1251
+ {
1252
+ "epoch": 1.2430167597765363,
1253
+ "grad_norm": 1.3041682243347168,
1254
+ "learning_rate": 2.039707851781791e-05,
1255
+ "loss": 0.2644,
1256
+ "step": 890
1257
+ },
1258
+ {
1259
+ "epoch": 1.25,
1260
+ "grad_norm": 1.548088788986206,
1261
+ "learning_rate": 2.0289155171182986e-05,
1262
+ "loss": 0.2672,
1263
+ "step": 895
1264
+ },
1265
+ {
1266
+ "epoch": 1.2569832402234637,
1267
+ "grad_norm": 1.411430835723877,
1268
+ "learning_rate": 2.0180918233554e-05,
1269
+ "loss": 0.2895,
1270
+ "step": 900
1271
+ },
1272
+ {
1273
+ "epoch": 1.2639664804469275,
1274
+ "grad_norm": 1.131250262260437,
1275
+ "learning_rate": 2.007237412223726e-05,
1276
+ "loss": 0.2667,
1277
+ "step": 905
1278
+ },
1279
+ {
1280
+ "epoch": 1.270949720670391,
1281
+ "grad_norm": 1.4170665740966797,
1282
+ "learning_rate": 1.9963529272751218e-05,
1283
+ "loss": 0.2756,
1284
+ "step": 910
1285
+ },
1286
+ {
1287
+ "epoch": 1.2779329608938548,
1288
+ "grad_norm": 1.5073738098144531,
1289
+ "learning_rate": 1.9854390138444957e-05,
1290
+ "loss": 0.2649,
1291
+ "step": 915
1292
+ },
1293
+ {
1294
+ "epoch": 1.2849162011173183,
1295
+ "grad_norm": 1.2252780199050903,
1296
+ "learning_rate": 1.97449631901155e-05,
1297
+ "loss": 0.2608,
1298
+ "step": 920
1299
+ },
1300
+ {
1301
+ "epoch": 1.291899441340782,
1302
+ "grad_norm": 1.1035939455032349,
1303
+ "learning_rate": 1.963525491562421e-05,
1304
+ "loss": 0.2509,
1305
+ "step": 925
1306
+ },
1307
+ {
1308
+ "epoch": 1.2988826815642458,
1309
+ "grad_norm": 1.1849076747894287,
1310
+ "learning_rate": 1.952527181951214e-05,
1311
+ "loss": 0.2719,
1312
+ "step": 930
1313
+ },
1314
+ {
1315
+ "epoch": 1.3058659217877095,
1316
+ "grad_norm": 1.3494380712509155,
1317
+ "learning_rate": 1.941502042261433e-05,
1318
+ "loss": 0.2805,
1319
+ "step": 935
1320
+ },
1321
+ {
1322
+ "epoch": 1.3128491620111733,
1323
+ "grad_norm": 1.2590703964233398,
1324
+ "learning_rate": 1.9304507261673238e-05,
1325
+ "loss": 0.2488,
1326
+ "step": 940
1327
+ },
1328
+ {
1329
+ "epoch": 1.3198324022346368,
1330
+ "grad_norm": 1.1709249019622803,
1331
+ "learning_rate": 1.919373888895116e-05,
1332
+ "loss": 0.264,
1333
+ "step": 945
1334
+ },
1335
+ {
1336
+ "epoch": 1.3268156424581006,
1337
+ "grad_norm": 1.2007285356521606,
1338
+ "learning_rate": 1.9082721871841746e-05,
1339
+ "loss": 0.2592,
1340
+ "step": 950
1341
+ },
1342
+ {
1343
+ "epoch": 1.3337988826815643,
1344
+ "grad_norm": 1.2193608283996582,
1345
+ "learning_rate": 1.8971462792480636e-05,
1346
+ "loss": 0.2526,
1347
+ "step": 955
1348
+ },
1349
+ {
1350
+ "epoch": 1.3407821229050279,
1351
+ "grad_norm": 1.1320077180862427,
1352
+ "learning_rate": 1.8859968247355208e-05,
1353
+ "loss": 0.2408,
1354
+ "step": 960
1355
+ },
1356
+ {
1357
+ "epoch": 1.3477653631284916,
1358
+ "grad_norm": 1.1588561534881592,
1359
+ "learning_rate": 1.8748244846913463e-05,
1360
+ "loss": 0.2412,
1361
+ "step": 965
1362
+ },
1363
+ {
1364
+ "epoch": 1.3547486033519553,
1365
+ "grad_norm": 1.5395079851150513,
1366
+ "learning_rate": 1.8636299215172102e-05,
1367
+ "loss": 0.2218,
1368
+ "step": 970
1369
+ },
1370
+ {
1371
+ "epoch": 1.361731843575419,
1372
+ "grad_norm": 1.1714344024658203,
1373
+ "learning_rate": 1.85241379893238e-05,
1374
+ "loss": 0.2526,
1375
+ "step": 975
1376
+ },
1377
+ {
1378
+ "epoch": 1.3687150837988826,
1379
+ "grad_norm": 1.0737510919570923,
1380
+ "learning_rate": 1.8411767819343672e-05,
1381
+ "loss": 0.2244,
1382
+ "step": 980
1383
+ },
1384
+ {
1385
+ "epoch": 1.3756983240223464,
1386
+ "grad_norm": 1.2870091199874878,
1387
+ "learning_rate": 1.8299195367595032e-05,
1388
+ "loss": 0.2214,
1389
+ "step": 985
1390
+ },
1391
+ {
1392
+ "epoch": 1.3826815642458101,
1393
+ "grad_norm": 1.1814113855361938,
1394
+ "learning_rate": 1.818642730843434e-05,
1395
+ "loss": 0.2243,
1396
+ "step": 990
1397
+ },
1398
+ {
1399
+ "epoch": 1.3896648044692737,
1400
+ "grad_norm": 1.201887845993042,
1401
+ "learning_rate": 1.8073470327815527e-05,
1402
+ "loss": 0.2402,
1403
+ "step": 995
1404
+ },
1405
+ {
1406
+ "epoch": 1.3966480446927374,
1407
+ "grad_norm": 1.1632310152053833,
1408
+ "learning_rate": 1.796033112289356e-05,
1409
+ "loss": 0.2153,
1410
+ "step": 1000
1411
+ },
1412
+ {
1413
+ "epoch": 1.4036312849162011,
1414
+ "grad_norm": 1.1883548498153687,
1415
+ "learning_rate": 1.7847016401627388e-05,
1416
+ "loss": 0.2114,
1417
+ "step": 1005
1418
+ },
1419
+ {
1420
+ "epoch": 1.410614525139665,
1421
+ "grad_norm": 1.2469031810760498,
1422
+ "learning_rate": 1.7733532882382217e-05,
1423
+ "loss": 0.2494,
1424
+ "step": 1010
1425
+ },
1426
+ {
1427
+ "epoch": 1.4175977653631284,
1428
+ "grad_norm": 1.1725187301635742,
1429
+ "learning_rate": 1.7619887293531185e-05,
1430
+ "loss": 0.1985,
1431
+ "step": 1015
1432
+ },
1433
+ {
1434
+ "epoch": 1.4245810055865922,
1435
+ "grad_norm": 1.1682534217834473,
1436
+ "learning_rate": 1.7506086373056454e-05,
1437
+ "loss": 0.2135,
1438
+ "step": 1020
1439
+ },
1440
+ {
1441
+ "epoch": 1.431564245810056,
1442
+ "grad_norm": 1.090573787689209,
1443
+ "learning_rate": 1.7392136868149713e-05,
1444
+ "loss": 0.2131,
1445
+ "step": 1025
1446
+ },
1447
+ {
1448
+ "epoch": 1.4385474860335195,
1449
+ "grad_norm": 1.14024817943573,
1450
+ "learning_rate": 1.7278045534812127e-05,
1451
+ "loss": 0.2043,
1452
+ "step": 1030
1453
+ },
1454
+ {
1455
+ "epoch": 1.4455307262569832,
1456
+ "grad_norm": 1.1963393688201904,
1457
+ "learning_rate": 1.7163819137453788e-05,
1458
+ "loss": 0.2216,
1459
+ "step": 1035
1460
+ },
1461
+ {
1462
+ "epoch": 1.452513966480447,
1463
+ "grad_norm": 0.9623016715049744,
1464
+ "learning_rate": 1.7049464448492665e-05,
1465
+ "loss": 0.2225,
1466
+ "step": 1040
1467
+ },
1468
+ {
1469
+ "epoch": 1.4594972067039107,
1470
+ "grad_norm": 1.0864177942276,
1471
+ "learning_rate": 1.6934988247953053e-05,
1472
+ "loss": 0.2066,
1473
+ "step": 1045
1474
+ },
1475
+ {
1476
+ "epoch": 1.4664804469273742,
1477
+ "grad_norm": 1.547531247138977,
1478
+ "learning_rate": 1.68203973230636e-05,
1479
+ "loss": 0.236,
1480
+ "step": 1050
1481
+ },
1482
+ {
1483
+ "epoch": 1.473463687150838,
1484
+ "grad_norm": 1.281081199645996,
1485
+ "learning_rate": 1.6705698467854918e-05,
1486
+ "loss": 0.1999,
1487
+ "step": 1055
1488
+ },
1489
+ {
1490
+ "epoch": 1.4804469273743017,
1491
+ "grad_norm": 1.0426186323165894,
1492
+ "learning_rate": 1.6590898482756713e-05,
1493
+ "loss": 0.1839,
1494
+ "step": 1060
1495
+ },
1496
+ {
1497
+ "epoch": 1.4874301675977653,
1498
+ "grad_norm": 1.1222844123840332,
1499
+ "learning_rate": 1.647600417419464e-05,
1500
+ "loss": 0.1729,
1501
+ "step": 1065
1502
+ },
1503
+ {
1504
+ "epoch": 1.494413407821229,
1505
+ "grad_norm": 1.181368350982666,
1506
+ "learning_rate": 1.636102235418674e-05,
1507
+ "loss": 0.2225,
1508
+ "step": 1070
1509
+ },
1510
+ {
1511
+ "epoch": 1.5013966480446927,
1512
+ "grad_norm": 1.0337244272232056,
1513
+ "learning_rate": 1.624595983993956e-05,
1514
+ "loss": 0.193,
1515
+ "step": 1075
1516
+ },
1517
+ {
1518
+ "epoch": 1.5083798882681565,
1519
+ "grad_norm": 1.2003777027130127,
1520
+ "learning_rate": 1.6130823453443948e-05,
1521
+ "loss": 0.1941,
1522
+ "step": 1080
1523
+ },
1524
+ {
1525
+ "epoch": 1.5153631284916202,
1526
+ "grad_norm": 1.0884066820144653,
1527
+ "learning_rate": 1.6015620021070613e-05,
1528
+ "loss": 0.1925,
1529
+ "step": 1085
1530
+ },
1531
+ {
1532
+ "epoch": 1.5223463687150838,
1533
+ "grad_norm": 1.254709005355835,
1534
+ "learning_rate": 1.5900356373165367e-05,
1535
+ "loss": 0.1846,
1536
+ "step": 1090
1537
+ },
1538
+ {
1539
+ "epoch": 1.5293296089385475,
1540
+ "grad_norm": 1.0388215780258179,
1541
+ "learning_rate": 1.578503934364416e-05,
1542
+ "loss": 0.1803,
1543
+ "step": 1095
1544
+ },
1545
+ {
1546
+ "epoch": 1.536312849162011,
1547
+ "grad_norm": 1.017614722251892,
1548
+ "learning_rate": 1.5669675769587924e-05,
1549
+ "loss": 0.1953,
1550
+ "step": 1100
1551
+ },
1552
+ {
1553
+ "epoch": 1.5432960893854748,
1554
+ "grad_norm": 1.0426898002624512,
1555
+ "learning_rate": 1.5554272490837176e-05,
1556
+ "loss": 0.1915,
1557
+ "step": 1105
1558
+ },
1559
+ {
1560
+ "epoch": 1.5502793296089385,
1561
+ "grad_norm": 0.9931243658065796,
1562
+ "learning_rate": 1.543883634958651e-05,
1563
+ "loss": 0.177,
1564
+ "step": 1110
1565
+ },
1566
+ {
1567
+ "epoch": 1.5572625698324023,
1568
+ "grad_norm": 1.1874698400497437,
1569
+ "learning_rate": 1.532337418997891e-05,
1570
+ "loss": 0.2084,
1571
+ "step": 1115
1572
+ },
1573
+ {
1574
+ "epoch": 1.564245810055866,
1575
+ "grad_norm": 1.1437238454818726,
1576
+ "learning_rate": 1.520789285769999e-05,
1577
+ "loss": 0.2002,
1578
+ "step": 1120
1579
+ },
1580
+ {
1581
+ "epoch": 1.5712290502793296,
1582
+ "grad_norm": 1.3141568899154663,
1583
+ "learning_rate": 1.5092399199572083e-05,
1584
+ "loss": 0.1879,
1585
+ "step": 1125
1586
+ },
1587
+ {
1588
+ "epoch": 1.5782122905027933,
1589
+ "grad_norm": 1.0645033121109009,
1590
+ "learning_rate": 1.497690006314832e-05,
1591
+ "loss": 0.193,
1592
+ "step": 1130
1593
+ },
1594
+ {
1595
+ "epoch": 1.5851955307262569,
1596
+ "grad_norm": 1.2224258184432983,
1597
+ "learning_rate": 1.486140229630664e-05,
1598
+ "loss": 0.1753,
1599
+ "step": 1135
1600
+ },
1601
+ {
1602
+ "epoch": 1.5921787709497206,
1603
+ "grad_norm": 1.3047986030578613,
1604
+ "learning_rate": 1.474591274684378e-05,
1605
+ "loss": 0.2036,
1606
+ "step": 1140
1607
+ },
1608
+ {
1609
+ "epoch": 1.5991620111731844,
1610
+ "grad_norm": 1.2557255029678345,
1611
+ "learning_rate": 1.4630438262069274e-05,
1612
+ "loss": 0.1808,
1613
+ "step": 1145
1614
+ },
1615
+ {
1616
+ "epoch": 1.606145251396648,
1617
+ "grad_norm": 1.195636510848999,
1618
+ "learning_rate": 1.4514985688399473e-05,
1619
+ "loss": 0.1902,
1620
+ "step": 1150
1621
+ },
1622
+ {
1623
+ "epoch": 1.6131284916201118,
1624
+ "grad_norm": 1.08878755569458,
1625
+ "learning_rate": 1.4399561870951637e-05,
1626
+ "loss": 0.1946,
1627
+ "step": 1155
1628
+ },
1629
+ {
1630
+ "epoch": 1.6201117318435754,
1631
+ "grad_norm": 1.2313425540924072,
1632
+ "learning_rate": 1.4284173653138085e-05,
1633
+ "loss": 0.1845,
1634
+ "step": 1160
1635
+ },
1636
+ {
1637
+ "epoch": 1.6270949720670391,
1638
+ "grad_norm": 1.0601774454116821,
1639
+ "learning_rate": 1.416882787626045e-05,
1640
+ "loss": 0.1873,
1641
+ "step": 1165
1642
+ },
1643
+ {
1644
+ "epoch": 1.6340782122905027,
1645
+ "grad_norm": 1.0548855066299438,
1646
+ "learning_rate": 1.4053531379104078e-05,
1647
+ "loss": 0.1966,
1648
+ "step": 1170
1649
+ },
1650
+ {
1651
+ "epoch": 1.6410614525139664,
1652
+ "grad_norm": 1.0141369104385376,
1653
+ "learning_rate": 1.3938290997532532e-05,
1654
+ "loss": 0.1869,
1655
+ "step": 1175
1656
+ },
1657
+ {
1658
+ "epoch": 1.6480446927374302,
1659
+ "grad_norm": 1.1447274684906006,
1660
+ "learning_rate": 1.3823113564082329e-05,
1661
+ "loss": 0.1901,
1662
+ "step": 1180
1663
+ },
1664
+ {
1665
+ "epoch": 1.655027932960894,
1666
+ "grad_norm": 1.0755653381347656,
1667
+ "learning_rate": 1.3708005907557812e-05,
1668
+ "loss": 0.1678,
1669
+ "step": 1185
1670
+ },
1671
+ {
1672
+ "epoch": 1.6620111731843576,
1673
+ "grad_norm": 1.2326295375823975,
1674
+ "learning_rate": 1.3592974852626303e-05,
1675
+ "loss": 0.1446,
1676
+ "step": 1190
1677
+ },
1678
+ {
1679
+ "epoch": 1.6689944134078212,
1680
+ "grad_norm": 1.1295117139816284,
1681
+ "learning_rate": 1.3478027219413457e-05,
1682
+ "loss": 0.168,
1683
+ "step": 1195
1684
+ },
1685
+ {
1686
+ "epoch": 1.675977653631285,
1687
+ "grad_norm": 1.0786888599395752,
1688
+ "learning_rate": 1.3363169823098898e-05,
1689
+ "loss": 0.1704,
1690
+ "step": 1200
1691
+ },
1692
+ {
1693
+ "epoch": 1.6829608938547485,
1694
+ "grad_norm": 1.2451069355010986,
1695
+ "learning_rate": 1.3248409473512158e-05,
1696
+ "loss": 0.1473,
1697
+ "step": 1205
1698
+ },
1699
+ {
1700
+ "epoch": 1.6899441340782122,
1701
+ "grad_norm": 1.251922607421875,
1702
+ "learning_rate": 1.3133752974728925e-05,
1703
+ "loss": 0.1736,
1704
+ "step": 1210
1705
+ },
1706
+ {
1707
+ "epoch": 1.696927374301676,
1708
+ "grad_norm": 1.0417166948318481,
1709
+ "learning_rate": 1.3019207124667632e-05,
1710
+ "loss": 0.1726,
1711
+ "step": 1215
1712
+ },
1713
+ {
1714
+ "epoch": 1.7039106145251397,
1715
+ "grad_norm": 1.092808723449707,
1716
+ "learning_rate": 1.2904778714686411e-05,
1717
+ "loss": 0.1633,
1718
+ "step": 1220
1719
+ },
1720
+ {
1721
+ "epoch": 1.7108938547486034,
1722
+ "grad_norm": 1.0576175451278687,
1723
+ "learning_rate": 1.2790474529180435e-05,
1724
+ "loss": 0.1611,
1725
+ "step": 1225
1726
+ },
1727
+ {
1728
+ "epoch": 1.7178770949720672,
1729
+ "grad_norm": 1.188326120376587,
1730
+ "learning_rate": 1.2676301345179692e-05,
1731
+ "loss": 0.1808,
1732
+ "step": 1230
1733
+ },
1734
+ {
1735
+ "epoch": 1.7248603351955307,
1736
+ "grad_norm": 1.0684913396835327,
1737
+ "learning_rate": 1.2562265931947154e-05,
1738
+ "loss": 0.1646,
1739
+ "step": 1235
1740
+ },
1741
+ {
1742
+ "epoch": 1.7318435754189943,
1743
+ "grad_norm": 1.4066827297210693,
1744
+ "learning_rate": 1.2448375050577441e-05,
1745
+ "loss": 0.1617,
1746
+ "step": 1240
1747
+ },
1748
+ {
1749
+ "epoch": 1.738826815642458,
1750
+ "grad_norm": 1.3388019800186157,
1751
+ "learning_rate": 1.233463545359597e-05,
1752
+ "loss": 0.1699,
1753
+ "step": 1245
1754
+ },
1755
+ {
1756
+ "epoch": 1.7458100558659218,
1757
+ "grad_norm": 1.1621959209442139,
1758
+ "learning_rate": 1.2221053884558586e-05,
1759
+ "loss": 0.1463,
1760
+ "step": 1250
1761
+ },
1762
+ {
1763
+ "epoch": 1.7527932960893855,
1764
+ "grad_norm": 1.1119213104248047,
1765
+ "learning_rate": 1.2107637077651749e-05,
1766
+ "loss": 0.1477,
1767
+ "step": 1255
1768
+ },
1769
+ {
1770
+ "epoch": 1.7597765363128492,
1771
+ "grad_norm": 1.2606251239776611,
1772
+ "learning_rate": 1.1994391757293271e-05,
1773
+ "loss": 0.1647,
1774
+ "step": 1260
1775
+ },
1776
+ {
1777
+ "epoch": 1.766759776536313,
1778
+ "grad_norm": 1.0008996725082397,
1779
+ "learning_rate": 1.1881324637733613e-05,
1780
+ "loss": 0.1679,
1781
+ "step": 1265
1782
+ },
1783
+ {
1784
+ "epoch": 1.7737430167597765,
1785
+ "grad_norm": 1.0280311107635498,
1786
+ "learning_rate": 1.1768442422657825e-05,
1787
+ "loss": 0.1538,
1788
+ "step": 1270
1789
+ },
1790
+ {
1791
+ "epoch": 1.7807262569832403,
1792
+ "grad_norm": 1.1013814210891724,
1793
+ "learning_rate": 1.1655751804788067e-05,
1794
+ "loss": 0.145,
1795
+ "step": 1275
1796
+ },
1797
+ {
1798
+ "epoch": 1.7877094972067038,
1799
+ "grad_norm": 0.9577712416648865,
1800
+ "learning_rate": 1.1543259465486817e-05,
1801
+ "loss": 0.1469,
1802
+ "step": 1280
1803
+ },
1804
+ {
1805
+ "epoch": 1.7946927374301676,
1806
+ "grad_norm": 1.0311791896820068,
1807
+ "learning_rate": 1.1430972074360722e-05,
1808
+ "loss": 0.1482,
1809
+ "step": 1285
1810
+ },
1811
+ {
1812
+ "epoch": 1.8016759776536313,
1813
+ "grad_norm": 1.1357600688934326,
1814
+ "learning_rate": 1.1318896288865169e-05,
1815
+ "loss": 0.1482,
1816
+ "step": 1290
1817
+ },
1818
+ {
1819
+ "epoch": 1.808659217877095,
1820
+ "grad_norm": 1.0834394693374634,
1821
+ "learning_rate": 1.120703875390958e-05,
1822
+ "loss": 0.1606,
1823
+ "step": 1295
1824
+ },
1825
+ {
1826
+ "epoch": 1.8156424581005588,
1827
+ "grad_norm": 1.1400260925292969,
1828
+ "learning_rate": 1.1095406101463414e-05,
1829
+ "loss": 0.1405,
1830
+ "step": 1300
1831
+ },
1832
+ {
1833
+ "epoch": 1.8226256983240223,
1834
+ "grad_norm": 1.101902723312378,
1835
+ "learning_rate": 1.098400495016298e-05,
1836
+ "loss": 0.1548,
1837
+ "step": 1305
1838
+ },
1839
+ {
1840
+ "epoch": 1.829608938547486,
1841
+ "grad_norm": 0.86611008644104,
1842
+ "learning_rate": 1.0872841904919012e-05,
1843
+ "loss": 0.1432,
1844
+ "step": 1310
1845
+ },
1846
+ {
1847
+ "epoch": 1.8365921787709496,
1848
+ "grad_norm": 1.0471363067626953,
1849
+ "learning_rate": 1.0761923556525074e-05,
1850
+ "loss": 0.1447,
1851
+ "step": 1315
1852
+ },
1853
+ {
1854
+ "epoch": 1.8435754189944134,
1855
+ "grad_norm": 1.162113070487976,
1856
+ "learning_rate": 1.0651256481266794e-05,
1857
+ "loss": 0.142,
1858
+ "step": 1320
1859
+ },
1860
+ {
1861
+ "epoch": 1.850558659217877,
1862
+ "grad_norm": 0.9480882883071899,
1863
+ "learning_rate": 1.054084724053196e-05,
1864
+ "loss": 0.1452,
1865
+ "step": 1325
1866
+ },
1867
+ {
1868
+ "epoch": 1.8575418994413408,
1869
+ "grad_norm": 0.9358695149421692,
1870
+ "learning_rate": 1.0430702380421496e-05,
1871
+ "loss": 0.137,
1872
+ "step": 1330
1873
+ },
1874
+ {
1875
+ "epoch": 1.8645251396648046,
1876
+ "grad_norm": 0.9968928694725037,
1877
+ "learning_rate": 1.0320828431361347e-05,
1878
+ "loss": 0.1457,
1879
+ "step": 1335
1880
+ },
1881
+ {
1882
+ "epoch": 1.8715083798882681,
1883
+ "grad_norm": 1.0149060487747192,
1884
+ "learning_rate": 1.0211231907715302e-05,
1885
+ "loss": 0.1349,
1886
+ "step": 1340
1887
+ },
1888
+ {
1889
+ "epoch": 1.8784916201117319,
1890
+ "grad_norm": 1.0668036937713623,
1891
+ "learning_rate": 1.0101919307398752e-05,
1892
+ "loss": 0.1242,
1893
+ "step": 1345
1894
+ },
1895
+ {
1896
+ "epoch": 1.8854748603351954,
1897
+ "grad_norm": 1.0963325500488281,
1898
+ "learning_rate": 9.992897111493438e-06,
1899
+ "loss": 0.1204,
1900
+ "step": 1350
1901
+ },
1902
+ {
1903
+ "epoch": 1.8924581005586592,
1904
+ "grad_norm": 0.9292417764663696,
1905
+ "learning_rate": 9.884171783863183e-06,
1906
+ "loss": 0.127,
1907
+ "step": 1355
1908
+ },
1909
+ {
1910
+ "epoch": 1.899441340782123,
1911
+ "grad_norm": 1.0576997995376587,
1912
+ "learning_rate": 9.775749770770672e-06,
1913
+ "loss": 0.1317,
1914
+ "step": 1360
1915
+ },
1916
+ {
1917
+ "epoch": 1.9064245810055866,
1918
+ "grad_norm": 1.055415391921997,
1919
+ "learning_rate": 9.66763750049523e-06,
1920
+ "loss": 0.1268,
1921
+ "step": 1365
1922
+ },
1923
+ {
1924
+ "epoch": 1.9134078212290504,
1925
+ "grad_norm": 1.01954984664917,
1926
+ "learning_rate": 9.559841382951713e-06,
1927
+ "loss": 0.1347,
1928
+ "step": 1370
1929
+ },
1930
+ {
1931
+ "epoch": 1.920391061452514,
1932
+ "grad_norm": 1.1383157968521118,
1933
+ "learning_rate": 9.452367809310463e-06,
1934
+ "loss": 0.1364,
1935
+ "step": 1375
1936
+ },
1937
+ {
1938
+ "epoch": 1.9273743016759777,
1939
+ "grad_norm": 0.9481402635574341,
1940
+ "learning_rate": 9.345223151618377e-06,
1941
+ "loss": 0.1252,
1942
+ "step": 1380
1943
+ },
1944
+ {
1945
+ "epoch": 1.9343575418994412,
1946
+ "grad_norm": 0.9682475924491882,
1947
+ "learning_rate": 9.238413762421114e-06,
1948
+ "loss": 0.1386,
1949
+ "step": 1385
1950
+ },
1951
+ {
1952
+ "epoch": 1.941340782122905,
1953
+ "grad_norm": 0.9600017666816711,
1954
+ "learning_rate": 9.13194597438645e-06,
1955
+ "loss": 0.122,
1956
+ "step": 1390
1957
+ },
1958
+ {
1959
+ "epoch": 1.9483240223463687,
1960
+ "grad_norm": 1.0056228637695312,
1961
+ "learning_rate": 9.025826099928826e-06,
1962
+ "loss": 0.1188,
1963
+ "step": 1395
1964
+ },
1965
+ {
1966
+ "epoch": 1.9553072625698324,
1967
+ "grad_norm": 0.8716211915016174,
1968
+ "learning_rate": 8.920060430835089e-06,
1969
+ "loss": 0.1176,
1970
+ "step": 1400
1971
+ },
1972
+ {
1973
+ "epoch": 1.9622905027932962,
1974
+ "grad_norm": 1.1095727682113647,
1975
+ "learning_rate": 8.81465523789145e-06,
1976
+ "loss": 0.1306,
1977
+ "step": 1405
1978
+ },
1979
+ {
1980
+ "epoch": 1.9692737430167597,
1981
+ "grad_norm": 0.83429354429245,
1982
+ "learning_rate": 8.70961677051169e-06,
1983
+ "loss": 0.1237,
1984
+ "step": 1410
1985
+ },
1986
+ {
1987
+ "epoch": 1.9762569832402235,
1988
+ "grad_norm": 0.992759108543396,
1989
+ "learning_rate": 8.604951256366648e-06,
1990
+ "loss": 0.1342,
1991
+ "step": 1415
1992
+ },
1993
+ {
1994
+ "epoch": 1.983240223463687,
1995
+ "grad_norm": 0.9825895428657532,
1996
+ "learning_rate": 8.50066490101496e-06,
1997
+ "loss": 0.1302,
1998
+ "step": 1420
1999
+ },
2000
+ {
2001
+ "epoch": 1.9902234636871508,
2002
+ "grad_norm": 1.0553414821624756,
2003
+ "learning_rate": 8.39676388753519e-06,
2004
+ "loss": 0.1177,
2005
+ "step": 1425
2006
+ },
2007
+ {
2008
+ "epoch": 1.9972067039106145,
2009
+ "grad_norm": 1.0035676956176758,
2010
+ "learning_rate": 8.293254376159165e-06,
2011
+ "loss": 0.1242,
2012
+ "step": 1430
2013
+ },
2014
+ {
2015
+ "epoch": 2.0041899441340782,
2016
+ "grad_norm": 0.8561891913414001,
2017
+ "learning_rate": 8.190142503906798e-06,
2018
+ "loss": 0.0949,
2019
+ "step": 1435
2020
+ },
2021
+ {
2022
+ "epoch": 2.011173184357542,
2023
+ "grad_norm": 0.9547761082649231,
2024
+ "learning_rate": 8.087434384222191e-06,
2025
+ "loss": 0.0991,
2026
+ "step": 1440
2027
+ },
2028
+ {
2029
+ "epoch": 2.0181564245810057,
2030
+ "grad_norm": 1.0174782276153564,
2031
+ "learning_rate": 7.9851361066112e-06,
2032
+ "loss": 0.0926,
2033
+ "step": 1445
2034
+ },
2035
+ {
2036
+ "epoch": 2.0251396648044695,
2037
+ "grad_norm": 0.9323471784591675,
2038
+ "learning_rate": 7.883253736280361e-06,
2039
+ "loss": 0.095,
2040
+ "step": 1450
2041
+ },
2042
+ {
2043
+ "epoch": 2.032122905027933,
2044
+ "grad_norm": 0.9076278209686279,
2045
+ "learning_rate": 7.781793313777326e-06,
2046
+ "loss": 0.0908,
2047
+ "step": 1455
2048
+ },
2049
+ {
2050
+ "epoch": 2.0391061452513966,
2051
+ "grad_norm": 1.2702761888504028,
2052
+ "learning_rate": 7.680760854632678e-06,
2053
+ "loss": 0.0949,
2054
+ "step": 1460
2055
+ },
2056
+ {
2057
+ "epoch": 2.0460893854748603,
2058
+ "grad_norm": 0.9120989441871643,
2059
+ "learning_rate": 7.580162349003321e-06,
2060
+ "loss": 0.1051,
2061
+ "step": 1465
2062
+ },
2063
+ {
2064
+ "epoch": 2.053072625698324,
2065
+ "grad_norm": 0.9788500070571899,
2066
+ "learning_rate": 7.480003761317275e-06,
2067
+ "loss": 0.0936,
2068
+ "step": 1470
2069
+ },
2070
+ {
2071
+ "epoch": 2.060055865921788,
2072
+ "grad_norm": 0.9200281500816345,
2073
+ "learning_rate": 7.380291029920105e-06,
2074
+ "loss": 0.0899,
2075
+ "step": 1475
2076
+ },
2077
+ {
2078
+ "epoch": 2.0670391061452515,
2079
+ "grad_norm": 0.795583963394165,
2080
+ "learning_rate": 7.281030066722782e-06,
2081
+ "loss": 0.094,
2082
+ "step": 1480
2083
+ },
2084
+ {
2085
+ "epoch": 2.0740223463687153,
2086
+ "grad_norm": 0.6631631255149841,
2087
+ "learning_rate": 7.182226756851223e-06,
2088
+ "loss": 0.0838,
2089
+ "step": 1485
2090
+ },
2091
+ {
2092
+ "epoch": 2.0810055865921786,
2093
+ "grad_norm": 0.8172409534454346,
2094
+ "learning_rate": 7.083886958297329e-06,
2095
+ "loss": 0.0881,
2096
+ "step": 1490
2097
+ },
2098
+ {
2099
+ "epoch": 2.0879888268156424,
2100
+ "grad_norm": 0.7226603031158447,
2101
+ "learning_rate": 6.9860165015716745e-06,
2102
+ "loss": 0.088,
2103
+ "step": 1495
2104
+ },
2105
+ {
2106
+ "epoch": 2.094972067039106,
2107
+ "grad_norm": 1.2240625619888306,
2108
+ "learning_rate": 6.8886211893578385e-06,
2109
+ "loss": 0.1011,
2110
+ "step": 1500
2111
+ },
2112
+ {
2113
+ "epoch": 2.10195530726257,
2114
+ "grad_norm": 0.7928593158721924,
2115
+ "learning_rate": 6.791706796168333e-06,
2116
+ "loss": 0.0906,
2117
+ "step": 1505
2118
+ },
2119
+ {
2120
+ "epoch": 2.1089385474860336,
2121
+ "grad_norm": 0.9772520661354065,
2122
+ "learning_rate": 6.695279068002278e-06,
2123
+ "loss": 0.0858,
2124
+ "step": 1510
2125
+ },
2126
+ {
2127
+ "epoch": 2.1159217877094973,
2128
+ "grad_norm": 0.8721750378608704,
2129
+ "learning_rate": 6.5993437220046795e-06,
2130
+ "loss": 0.1038,
2131
+ "step": 1515
2132
+ },
2133
+ {
2134
+ "epoch": 2.122905027932961,
2135
+ "grad_norm": 0.7537911534309387,
2136
+ "learning_rate": 6.503906446127511e-06,
2137
+ "loss": 0.0991,
2138
+ "step": 1520
2139
+ },
2140
+ {
2141
+ "epoch": 2.1298882681564244,
2142
+ "grad_norm": 0.8837329745292664,
2143
+ "learning_rate": 6.408972898792423e-06,
2144
+ "loss": 0.0916,
2145
+ "step": 1525
2146
+ },
2147
+ {
2148
+ "epoch": 2.136871508379888,
2149
+ "grad_norm": 0.8966924548149109,
2150
+ "learning_rate": 6.314548708555315e-06,
2151
+ "loss": 0.0895,
2152
+ "step": 1530
2153
+ },
2154
+ {
2155
+ "epoch": 2.143854748603352,
2156
+ "grad_norm": 1.077112078666687,
2157
+ "learning_rate": 6.2206394737725725e-06,
2158
+ "loss": 0.1007,
2159
+ "step": 1535
2160
+ },
2161
+ {
2162
+ "epoch": 2.1508379888268156,
2163
+ "grad_norm": 0.8577045202255249,
2164
+ "learning_rate": 6.127250762269186e-06,
2165
+ "loss": 0.0875,
2166
+ "step": 1540
2167
+ },
2168
+ {
2169
+ "epoch": 2.1578212290502794,
2170
+ "grad_norm": 0.7255130410194397,
2171
+ "learning_rate": 6.0343881110085965e-06,
2172
+ "loss": 0.0826,
2173
+ "step": 1545
2174
+ },
2175
+ {
2176
+ "epoch": 2.164804469273743,
2177
+ "grad_norm": 0.8676854968070984,
2178
+ "learning_rate": 5.942057025764452e-06,
2179
+ "loss": 0.0878,
2180
+ "step": 1550
2181
+ },
2182
+ {
2183
+ "epoch": 2.171787709497207,
2184
+ "grad_norm": 0.7912610173225403,
2185
+ "learning_rate": 5.85026298079415e-06,
2186
+ "loss": 0.0992,
2187
+ "step": 1555
2188
+ },
2189
+ {
2190
+ "epoch": 2.17877094972067,
2191
+ "grad_norm": 0.9239996075630188,
2192
+ "learning_rate": 5.7590114185142625e-06,
2193
+ "loss": 0.095,
2194
+ "step": 1560
2195
+ },
2196
+ {
2197
+ "epoch": 2.185754189944134,
2198
+ "grad_norm": 0.8218305706977844,
2199
+ "learning_rate": 5.6683077491778935e-06,
2200
+ "loss": 0.0911,
2201
+ "step": 1565
2202
+ },
2203
+ {
2204
+ "epoch": 2.1927374301675977,
2205
+ "grad_norm": 1.0204100608825684,
2206
+ "learning_rate": 5.578157350553867e-06,
2207
+ "loss": 0.085,
2208
+ "step": 1570
2209
+ },
2210
+ {
2211
+ "epoch": 2.1997206703910615,
2212
+ "grad_norm": 0.7896533012390137,
2213
+ "learning_rate": 5.48856556760792e-06,
2214
+ "loss": 0.0801,
2215
+ "step": 1575
2216
+ },
2217
+ {
2218
+ "epoch": 2.206703910614525,
2219
+ "grad_norm": 0.7958258390426636,
2220
+ "learning_rate": 5.399537712185764e-06,
2221
+ "loss": 0.0773,
2222
+ "step": 1580
2223
+ },
2224
+ {
2225
+ "epoch": 2.213687150837989,
2226
+ "grad_norm": 0.6316238045692444,
2227
+ "learning_rate": 5.3110790626981935e-06,
2228
+ "loss": 0.0799,
2229
+ "step": 1585
2230
+ },
2231
+ {
2232
+ "epoch": 2.2206703910614527,
2233
+ "grad_norm": 0.8786280751228333,
2234
+ "learning_rate": 5.223194863808084e-06,
2235
+ "loss": 0.0896,
2236
+ "step": 1590
2237
+ },
2238
+ {
2239
+ "epoch": 2.227653631284916,
2240
+ "grad_norm": 0.8636316657066345,
2241
+ "learning_rate": 5.135890326119482e-06,
2242
+ "loss": 0.0962,
2243
+ "step": 1595
2244
+ },
2245
+ {
2246
+ "epoch": 2.2346368715083798,
2247
+ "grad_norm": 0.7825942635536194,
2248
+ "learning_rate": 5.049170625868636e-06,
2249
+ "loss": 0.0879,
2250
+ "step": 1600
2251
+ },
2252
+ {
2253
+ "epoch": 2.2416201117318435,
2254
+ "grad_norm": 0.6920234560966492,
2255
+ "learning_rate": 4.963040904617131e-06,
2256
+ "loss": 0.0735,
2257
+ "step": 1605
2258
+ },
2259
+ {
2260
+ "epoch": 2.2486033519553073,
2261
+ "grad_norm": 1.11583411693573,
2262
+ "learning_rate": 4.877506268947014e-06,
2263
+ "loss": 0.0892,
2264
+ "step": 1610
2265
+ },
2266
+ {
2267
+ "epoch": 2.255586592178771,
2268
+ "grad_norm": 1.003596305847168,
2269
+ "learning_rate": 4.792571790158077e-06,
2270
+ "loss": 0.0958,
2271
+ "step": 1615
2272
+ },
2273
+ {
2274
+ "epoch": 2.2625698324022347,
2275
+ "grad_norm": 0.8565225005149841,
2276
+ "learning_rate": 4.708242503967122e-06,
2277
+ "loss": 0.0829,
2278
+ "step": 1620
2279
+ },
2280
+ {
2281
+ "epoch": 2.2695530726256985,
2282
+ "grad_norm": 0.815064549446106,
2283
+ "learning_rate": 4.62452341020945e-06,
2284
+ "loss": 0.0835,
2285
+ "step": 1625
2286
+ },
2287
+ {
2288
+ "epoch": 2.276536312849162,
2289
+ "grad_norm": 0.8153538703918457,
2290
+ "learning_rate": 4.54141947254238e-06,
2291
+ "loss": 0.0802,
2292
+ "step": 1630
2293
+ },
2294
+ {
2295
+ "epoch": 2.2835195530726256,
2296
+ "grad_norm": 0.903838574886322,
2297
+ "learning_rate": 4.458935618151e-06,
2298
+ "loss": 0.0802,
2299
+ "step": 1635
2300
+ },
2301
+ {
2302
+ "epoch": 2.2905027932960893,
2303
+ "grad_norm": 0.6654587984085083,
2304
+ "learning_rate": 4.377076737455994e-06,
2305
+ "loss": 0.077,
2306
+ "step": 1640
2307
+ },
2308
+ {
2309
+ "epoch": 2.297486033519553,
2310
+ "grad_norm": 0.7662801742553711,
2311
+ "learning_rate": 4.2958476838237165e-06,
2312
+ "loss": 0.0774,
2313
+ "step": 1645
2314
+ },
2315
+ {
2316
+ "epoch": 2.304469273743017,
2317
+ "grad_norm": 0.7813943028450012,
2318
+ "learning_rate": 4.215253273278449e-06,
2319
+ "loss": 0.0772,
2320
+ "step": 1650
2321
+ },
2322
+ {
2323
+ "epoch": 2.3114525139664805,
2324
+ "grad_norm": 0.6762244701385498,
2325
+ "learning_rate": 4.135298284216825e-06,
2326
+ "loss": 0.0825,
2327
+ "step": 1655
2328
+ },
2329
+ {
2330
+ "epoch": 2.3184357541899443,
2331
+ "grad_norm": 0.6521827578544617,
2332
+ "learning_rate": 4.055987457124563e-06,
2333
+ "loss": 0.0854,
2334
+ "step": 1660
2335
+ },
2336
+ {
2337
+ "epoch": 2.3254189944134076,
2338
+ "grad_norm": 0.6951796412467957,
2339
+ "learning_rate": 3.977325494295367e-06,
2340
+ "loss": 0.0831,
2341
+ "step": 1665
2342
+ },
2343
+ {
2344
+ "epoch": 2.3324022346368714,
2345
+ "grad_norm": 0.7081191539764404,
2346
+ "learning_rate": 3.899317059552165e-06,
2347
+ "loss": 0.0793,
2348
+ "step": 1670
2349
+ },
2350
+ {
2351
+ "epoch": 2.339385474860335,
2352
+ "grad_norm": 0.7401983141899109,
2353
+ "learning_rate": 3.82196677797056e-06,
2354
+ "loss": 0.0871,
2355
+ "step": 1675
2356
+ },
2357
+ {
2358
+ "epoch": 2.346368715083799,
2359
+ "grad_norm": 0.7391781210899353,
2360
+ "learning_rate": 3.745279235604653e-06,
2361
+ "loss": 0.0812,
2362
+ "step": 1680
2363
+ },
2364
+ {
2365
+ "epoch": 2.3533519553072626,
2366
+ "grad_norm": 0.8532117009162903,
2367
+ "learning_rate": 3.6692589792150923e-06,
2368
+ "loss": 0.0805,
2369
+ "step": 1685
2370
+ },
2371
+ {
2372
+ "epoch": 2.3603351955307263,
2373
+ "grad_norm": 0.6902256608009338,
2374
+ "learning_rate": 3.5939105159995363e-06,
2375
+ "loss": 0.0769,
2376
+ "step": 1690
2377
+ },
2378
+ {
2379
+ "epoch": 2.36731843575419,
2380
+ "grad_norm": 0.776439905166626,
2381
+ "learning_rate": 3.5192383133253975e-06,
2382
+ "loss": 0.0831,
2383
+ "step": 1695
2384
+ },
2385
+ {
2386
+ "epoch": 2.3743016759776534,
2387
+ "grad_norm": 0.9593327641487122,
2388
+ "learning_rate": 3.4452467984649967e-06,
2389
+ "loss": 0.0822,
2390
+ "step": 1700
2391
+ },
2392
+ {
2393
+ "epoch": 2.381284916201117,
2394
+ "grad_norm": 0.5840911269187927,
2395
+ "learning_rate": 3.3719403583330516e-06,
2396
+ "loss": 0.0743,
2397
+ "step": 1705
2398
+ },
2399
+ {
2400
+ "epoch": 2.388268156424581,
2401
+ "grad_norm": 0.670928418636322,
2402
+ "learning_rate": 3.2993233392266046e-06,
2403
+ "loss": 0.0823,
2404
+ "step": 1710
2405
+ },
2406
+ {
2407
+ "epoch": 2.3952513966480447,
2408
+ "grad_norm": 0.699812650680542,
2409
+ "learning_rate": 3.227400046567302e-06,
2410
+ "loss": 0.0845,
2411
+ "step": 1715
2412
+ },
2413
+ {
2414
+ "epoch": 2.4022346368715084,
2415
+ "grad_norm": 0.8536289930343628,
2416
+ "learning_rate": 3.156174744646162e-06,
2417
+ "loss": 0.0758,
2418
+ "step": 1720
2419
+ },
2420
+ {
2421
+ "epoch": 2.409217877094972,
2422
+ "grad_norm": 0.7806151509284973,
2423
+ "learning_rate": 3.08565165637071e-06,
2424
+ "loss": 0.0798,
2425
+ "step": 1725
2426
+ },
2427
+ {
2428
+ "epoch": 2.416201117318436,
2429
+ "grad_norm": 0.8904385566711426,
2430
+ "learning_rate": 3.015834963014643e-06,
2431
+ "loss": 0.0856,
2432
+ "step": 1730
2433
+ },
2434
+ {
2435
+ "epoch": 2.423184357541899,
2436
+ "grad_norm": 0.663139283657074,
2437
+ "learning_rate": 2.9467288039698943e-06,
2438
+ "loss": 0.0742,
2439
+ "step": 1735
2440
+ },
2441
+ {
2442
+ "epoch": 2.430167597765363,
2443
+ "grad_norm": 0.7719913721084595,
2444
+ "learning_rate": 2.878337276501229e-06,
2445
+ "loss": 0.0868,
2446
+ "step": 1740
2447
+ },
2448
+ {
2449
+ "epoch": 2.4371508379888267,
2450
+ "grad_norm": 0.5354697704315186,
2451
+ "learning_rate": 2.8106644355033175e-06,
2452
+ "loss": 0.0749,
2453
+ "step": 1745
2454
+ },
2455
+ {
2456
+ "epoch": 2.4441340782122905,
2457
+ "grad_norm": 0.6630569100379944,
2458
+ "learning_rate": 2.7437142932603027e-06,
2459
+ "loss": 0.0797,
2460
+ "step": 1750
2461
+ },
2462
+ {
2463
+ "epoch": 2.451117318435754,
2464
+ "grad_norm": 0.903147280216217,
2465
+ "learning_rate": 2.67749081920795e-06,
2466
+ "loss": 0.0783,
2467
+ "step": 1755
2468
+ },
2469
+ {
2470
+ "epoch": 2.458100558659218,
2471
+ "grad_norm": 0.6932404041290283,
2472
+ "learning_rate": 2.611997939698268e-06,
2473
+ "loss": 0.0854,
2474
+ "step": 1760
2475
+ },
2476
+ {
2477
+ "epoch": 2.4650837988826817,
2478
+ "grad_norm": 0.6541863679885864,
2479
+ "learning_rate": 2.547239537766743e-06,
2480
+ "loss": 0.0691,
2481
+ "step": 1765
2482
+ },
2483
+ {
2484
+ "epoch": 2.472067039106145,
2485
+ "grad_norm": 0.7833231687545776,
2486
+ "learning_rate": 2.4832194529020884e-06,
2487
+ "loss": 0.0695,
2488
+ "step": 1770
2489
+ },
2490
+ {
2491
+ "epoch": 2.4790502793296088,
2492
+ "grad_norm": 0.739987313747406,
2493
+ "learning_rate": 2.419941480818641e-06,
2494
+ "loss": 0.0742,
2495
+ "step": 1775
2496
+ },
2497
+ {
2498
+ "epoch": 2.4860335195530725,
2499
+ "grad_norm": 0.74677574634552,
2500
+ "learning_rate": 2.357409373231274e-06,
2501
+ "loss": 0.088,
2502
+ "step": 1780
2503
+ },
2504
+ {
2505
+ "epoch": 2.4930167597765363,
2506
+ "grad_norm": 0.7661157846450806,
2507
+ "learning_rate": 2.2956268376329953e-06,
2508
+ "loss": 0.0711,
2509
+ "step": 1785
2510
+ },
2511
+ {
2512
+ "epoch": 2.5,
2513
+ "grad_norm": 0.7488077282905579,
2514
+ "learning_rate": 2.234597537075104e-06,
2515
+ "loss": 0.0822,
2516
+ "step": 1790
2517
+ },
2518
+ {
2519
+ "epoch": 2.5069832402234637,
2520
+ "grad_norm": 0.6147714257240295,
2521
+ "learning_rate": 2.174325089950035e-06,
2522
+ "loss": 0.0773,
2523
+ "step": 1795
2524
+ },
2525
+ {
2526
+ "epoch": 2.5139664804469275,
2527
+ "grad_norm": 0.5596182346343994,
2528
+ "learning_rate": 2.114813069776805e-06,
2529
+ "loss": 0.0809,
2530
+ "step": 1800
2531
+ },
2532
+ {
2533
+ "epoch": 2.520949720670391,
2534
+ "grad_norm": 0.5971940159797668,
2535
+ "learning_rate": 2.056065004989155e-06,
2536
+ "loss": 0.071,
2537
+ "step": 1805
2538
+ },
2539
+ {
2540
+ "epoch": 2.527932960893855,
2541
+ "grad_norm": 0.6062336564064026,
2542
+ "learning_rate": 1.998084378726351e-06,
2543
+ "loss": 0.0713,
2544
+ "step": 1810
2545
+ },
2546
+ {
2547
+ "epoch": 2.5349162011173183,
2548
+ "grad_norm": 0.6327534317970276,
2549
+ "learning_rate": 1.940874628626655e-06,
2550
+ "loss": 0.082,
2551
+ "step": 1815
2552
+ },
2553
+ {
2554
+ "epoch": 2.541899441340782,
2555
+ "grad_norm": 0.5518718957901001,
2556
+ "learning_rate": 1.8844391466235361e-06,
2557
+ "loss": 0.0699,
2558
+ "step": 1820
2559
+ },
2560
+ {
2561
+ "epoch": 2.548882681564246,
2562
+ "grad_norm": 0.6850572824478149,
2563
+ "learning_rate": 1.8287812787445312e-06,
2564
+ "loss": 0.0743,
2565
+ "step": 1825
2566
+ },
2567
+ {
2568
+ "epoch": 2.5558659217877095,
2569
+ "grad_norm": 0.6397345662117004,
2570
+ "learning_rate": 1.7739043249128983e-06,
2571
+ "loss": 0.0783,
2572
+ "step": 1830
2573
+ },
2574
+ {
2575
+ "epoch": 2.5628491620111733,
2576
+ "grad_norm": 0.6533808708190918,
2577
+ "learning_rate": 1.7198115387519297e-06,
2578
+ "loss": 0.0759,
2579
+ "step": 1835
2580
+ },
2581
+ {
2582
+ "epoch": 2.5698324022346366,
2583
+ "grad_norm": 0.6530839204788208,
2584
+ "learning_rate": 1.66650612739208e-06,
2585
+ "loss": 0.0781,
2586
+ "step": 1840
2587
+ },
2588
+ {
2589
+ "epoch": 2.576815642458101,
2590
+ "grad_norm": 0.6948469877243042,
2591
+ "learning_rate": 1.613991251280783e-06,
2592
+ "loss": 0.0729,
2593
+ "step": 1845
2594
+ },
2595
+ {
2596
+ "epoch": 2.583798882681564,
2597
+ "grad_norm": 0.7088574767112732,
2598
+ "learning_rate": 1.5622700239951066e-06,
2599
+ "loss": 0.0653,
2600
+ "step": 1850
2601
+ },
2602
+ {
2603
+ "epoch": 2.590782122905028,
2604
+ "grad_norm": 0.5940538644790649,
2605
+ "learning_rate": 1.5113455120571169e-06,
2606
+ "loss": 0.0715,
2607
+ "step": 1855
2608
+ },
2609
+ {
2610
+ "epoch": 2.5977653631284916,
2611
+ "grad_norm": 0.6278888583183289,
2612
+ "learning_rate": 1.461220734752094e-06,
2613
+ "loss": 0.0766,
2614
+ "step": 1860
2615
+ },
2616
+ {
2617
+ "epoch": 2.6047486033519553,
2618
+ "grad_norm": 0.6526880264282227,
2619
+ "learning_rate": 1.4118986639494919e-06,
2620
+ "loss": 0.0753,
2621
+ "step": 1865
2622
+ },
2623
+ {
2624
+ "epoch": 2.611731843575419,
2625
+ "grad_norm": 0.7749740481376648,
2626
+ "learning_rate": 1.3633822239267757e-06,
2627
+ "loss": 0.0815,
2628
+ "step": 1870
2629
+ },
2630
+ {
2631
+ "epoch": 2.618715083798883,
2632
+ "grad_norm": 0.6862809658050537,
2633
+ "learning_rate": 1.3156742911960006e-06,
2634
+ "loss": 0.0769,
2635
+ "step": 1875
2636
+ },
2637
+ {
2638
+ "epoch": 2.6256983240223466,
2639
+ "grad_norm": 0.7089734077453613,
2640
+ "learning_rate": 1.2687776943333e-06,
2641
+ "loss": 0.0712,
2642
+ "step": 1880
2643
+ },
2644
+ {
2645
+ "epoch": 2.63268156424581,
2646
+ "grad_norm": 0.634204626083374,
2647
+ "learning_rate": 1.2226952138111546e-06,
2648
+ "loss": 0.0752,
2649
+ "step": 1885
2650
+ },
2651
+ {
2652
+ "epoch": 2.6396648044692737,
2653
+ "grad_norm": 0.5228337049484253,
2654
+ "learning_rate": 1.1774295818335607e-06,
2655
+ "loss": 0.0751,
2656
+ "step": 1890
2657
+ },
2658
+ {
2659
+ "epoch": 2.6466480446927374,
2660
+ "grad_norm": 0.6062474846839905,
2661
+ "learning_rate": 1.1329834821740231e-06,
2662
+ "loss": 0.0709,
2663
+ "step": 1895
2664
+ },
2665
+ {
2666
+ "epoch": 2.653631284916201,
2667
+ "grad_norm": 0.7614616751670837,
2668
+ "learning_rate": 1.0893595500164465e-06,
2669
+ "loss": 0.0795,
2670
+ "step": 1900
2671
+ },
2672
+ {
2673
+ "epoch": 2.660614525139665,
2674
+ "grad_norm": 0.5724729299545288,
2675
+ "learning_rate": 1.0465603717988982e-06,
2676
+ "loss": 0.071,
2677
+ "step": 1905
2678
+ },
2679
+ {
2680
+ "epoch": 2.6675977653631286,
2681
+ "grad_norm": 0.6923154592514038,
2682
+ "learning_rate": 1.0045884850602399e-06,
2683
+ "loss": 0.0712,
2684
+ "step": 1910
2685
+ },
2686
+ {
2687
+ "epoch": 2.6745810055865924,
2688
+ "grad_norm": 0.5562082529067993,
2689
+ "learning_rate": 9.634463782897135e-07,
2690
+ "loss": 0.0697,
2691
+ "step": 1915
2692
+ },
2693
+ {
2694
+ "epoch": 2.6815642458100557,
2695
+ "grad_norm": 0.5666680932044983,
2696
+ "learning_rate": 9.231364907793588e-07,
2697
+ "loss": 0.0737,
2698
+ "step": 1920
2699
+ },
2700
+ {
2701
+ "epoch": 2.6885474860335195,
2702
+ "grad_norm": 0.7212756276130676,
2703
+ "learning_rate": 8.836612124794285e-07,
2704
+ "loss": 0.0711,
2705
+ "step": 1925
2706
+ },
2707
+ {
2708
+ "epoch": 2.695530726256983,
2709
+ "grad_norm": 0.7551804780960083,
2710
+ "learning_rate": 8.450228838566549e-07,
2711
+ "loss": 0.0781,
2712
+ "step": 1930
2713
+ },
2714
+ {
2715
+ "epoch": 2.702513966480447,
2716
+ "grad_norm": 0.5251168012619019,
2717
+ "learning_rate": 8.0722379575552e-07,
2718
+ "loss": 0.0786,
2719
+ "step": 1935
2720
+ },
2721
+ {
2722
+ "epoch": 2.7094972067039107,
2723
+ "grad_norm": 0.6626659035682678,
2724
+ "learning_rate": 7.702661892623941e-07,
2725
+ "loss": 0.0651,
2726
+ "step": 1940
2727
+ },
2728
+ {
2729
+ "epoch": 2.7164804469273744,
2730
+ "grad_norm": 0.6480070352554321,
2731
+ "learning_rate": 7.341522555726971e-07,
2732
+ "loss": 0.0754,
2733
+ "step": 1945
2734
+ },
2735
+ {
2736
+ "epoch": 2.723463687150838,
2737
+ "grad_norm": 0.5670280456542969,
2738
+ "learning_rate": 6.988841358609599e-07,
2739
+ "loss": 0.0695,
2740
+ "step": 1950
2741
+ },
2742
+ {
2743
+ "epoch": 2.7304469273743015,
2744
+ "grad_norm": 0.6045445799827576,
2745
+ "learning_rate": 6.644639211538944e-07,
2746
+ "loss": 0.0751,
2747
+ "step": 1955
2748
+ },
2749
+ {
2750
+ "epoch": 2.7374301675977653,
2751
+ "grad_norm": 0.5695311427116394,
2752
+ "learning_rate": 6.308936522063991e-07,
2753
+ "loss": 0.0624,
2754
+ "step": 1960
2755
+ },
2756
+ {
2757
+ "epoch": 2.744413407821229,
2758
+ "grad_norm": 0.5639582872390747,
2759
+ "learning_rate": 5.981753193805789e-07,
2760
+ "loss": 0.0656,
2761
+ "step": 1965
2762
+ },
2763
+ {
2764
+ "epoch": 2.7513966480446927,
2765
+ "grad_norm": 0.5985326170921326,
2766
+ "learning_rate": 5.663108625277285e-07,
2767
+ "loss": 0.0725,
2768
+ "step": 1970
2769
+ },
2770
+ {
2771
+ "epoch": 2.7583798882681565,
2772
+ "grad_norm": 0.7429053783416748,
2773
+ "learning_rate": 5.353021708733274e-07,
2774
+ "loss": 0.0813,
2775
+ "step": 1975
2776
+ },
2777
+ {
2778
+ "epoch": 2.7653631284916202,
2779
+ "grad_norm": 0.5742406249046326,
2780
+ "learning_rate": 5.051510829050193e-07,
2781
+ "loss": 0.0687,
2782
+ "step": 1980
2783
+ },
2784
+ {
2785
+ "epoch": 2.772346368715084,
2786
+ "grad_norm": 0.5024329423904419,
2787
+ "learning_rate": 4.758593862636179e-07,
2788
+ "loss": 0.0723,
2789
+ "step": 1985
2790
+ },
2791
+ {
2792
+ "epoch": 2.7793296089385473,
2793
+ "grad_norm": 0.6796080470085144,
2794
+ "learning_rate": 4.474288176371133e-07,
2795
+ "loss": 0.0741,
2796
+ "step": 1990
2797
+ },
2798
+ {
2799
+ "epoch": 2.786312849162011,
2800
+ "grad_norm": 0.6728164553642273,
2801
+ "learning_rate": 4.1986106265770763e-07,
2802
+ "loss": 0.0652,
2803
+ "step": 1995
2804
+ },
2805
+ {
2806
+ "epoch": 2.793296089385475,
2807
+ "grad_norm": 0.6413192749023438,
2808
+ "learning_rate": 3.931577558018751e-07,
2809
+ "loss": 0.0808,
2810
+ "step": 2000
2811
+ },
2812
+ {
2813
+ "epoch": 2.8002793296089385,
2814
+ "grad_norm": 0.5070434212684631,
2815
+ "learning_rate": 3.67320480293451e-07,
2816
+ "loss": 0.0639,
2817
+ "step": 2005
2818
+ },
2819
+ {
2820
+ "epoch": 2.8072625698324023,
2821
+ "grad_norm": 0.6241849064826965,
2822
+ "learning_rate": 3.4235076800976706e-07,
2823
+ "loss": 0.0674,
2824
+ "step": 2010
2825
+ },
2826
+ {
2827
+ "epoch": 2.814245810055866,
2828
+ "grad_norm": 0.5576947927474976,
2829
+ "learning_rate": 3.182500993908255e-07,
2830
+ "loss": 0.0693,
2831
+ "step": 2015
2832
+ },
2833
+ {
2834
+ "epoch": 2.82122905027933,
2835
+ "grad_norm": 0.6266166567802429,
2836
+ "learning_rate": 2.9501990335152775e-07,
2837
+ "loss": 0.0718,
2838
+ "step": 2020
2839
+ },
2840
+ {
2841
+ "epoch": 2.828212290502793,
2842
+ "grad_norm": 0.7786450982093811,
2843
+ "learning_rate": 2.7266155719694706e-07,
2844
+ "loss": 0.0738,
2845
+ "step": 2025
2846
+ },
2847
+ {
2848
+ "epoch": 2.835195530726257,
2849
+ "grad_norm": 0.5644041299819946,
2850
+ "learning_rate": 2.5117638654068243e-07,
2851
+ "loss": 0.0656,
2852
+ "step": 2030
2853
+ },
2854
+ {
2855
+ "epoch": 2.8421787709497206,
2856
+ "grad_norm": 0.6355459690093994,
2857
+ "learning_rate": 2.305656652262461e-07,
2858
+ "loss": 0.0657,
2859
+ "step": 2035
2860
+ },
2861
+ {
2862
+ "epoch": 2.8491620111731844,
2863
+ "grad_norm": 0.5660210251808167,
2864
+ "learning_rate": 2.1083061525155433e-07,
2865
+ "loss": 0.0615,
2866
+ "step": 2040
2867
+ },
2868
+ {
2869
+ "epoch": 2.856145251396648,
2870
+ "grad_norm": 0.8288242220878601,
2871
+ "learning_rate": 1.919724066964651e-07,
2872
+ "loss": 0.0759,
2873
+ "step": 2045
2874
+ },
2875
+ {
2876
+ "epoch": 2.863128491620112,
2877
+ "grad_norm": 0.6449695825576782,
2878
+ "learning_rate": 1.7399215765341204e-07,
2879
+ "loss": 0.0728,
2880
+ "step": 2050
2881
+ },
2882
+ {
2883
+ "epoch": 2.8701117318435756,
2884
+ "grad_norm": 0.6770250797271729,
2885
+ "learning_rate": 1.5689093416110744e-07,
2886
+ "loss": 0.0769,
2887
+ "step": 2055
2888
+ },
2889
+ {
2890
+ "epoch": 2.877094972067039,
2891
+ "grad_norm": 0.5850970149040222,
2892
+ "learning_rate": 1.4066975014133787e-07,
2893
+ "loss": 0.0767,
2894
+ "step": 2060
2895
+ },
2896
+ {
2897
+ "epoch": 2.8840782122905027,
2898
+ "grad_norm": 0.7905807495117188,
2899
+ "learning_rate": 1.253295673388588e-07,
2900
+ "loss": 0.0757,
2901
+ "step": 2065
2902
+ },
2903
+ {
2904
+ "epoch": 2.8910614525139664,
2905
+ "grad_norm": 0.5646936297416687,
2906
+ "learning_rate": 1.1087129526435702e-07,
2907
+ "loss": 0.061,
2908
+ "step": 2070
2909
+ },
2910
+ {
2911
+ "epoch": 2.89804469273743,
2912
+ "grad_norm": 0.6773931980133057,
2913
+ "learning_rate": 9.729579114054366e-08,
2914
+ "loss": 0.0662,
2915
+ "step": 2075
2916
+ },
2917
+ {
2918
+ "epoch": 2.905027932960894,
2919
+ "grad_norm": 0.5431219339370728,
2920
+ "learning_rate": 8.460385985131159e-08,
2921
+ "loss": 0.0742,
2922
+ "step": 2080
2923
+ },
2924
+ {
2925
+ "epoch": 2.9120111731843576,
2926
+ "grad_norm": 0.8402865529060364,
2927
+ "learning_rate": 7.279625389403355e-08,
2928
+ "loss": 0.0739,
2929
+ "step": 2085
2930
+ },
2931
+ {
2932
+ "epoch": 2.9189944134078214,
2933
+ "grad_norm": 0.6484595537185669,
2934
+ "learning_rate": 6.187367333493288e-08,
2935
+ "loss": 0.0819,
2936
+ "step": 2090
2937
+ },
2938
+ {
2939
+ "epoch": 2.9259776536312847,
2940
+ "grad_norm": 0.7217246890068054,
2941
+ "learning_rate": 5.183676576758167e-08,
2942
+ "loss": 0.0702,
2943
+ "step": 2095
2944
+ },
2945
+ {
2946
+ "epoch": 2.9329608938547485,
2947
+ "grad_norm": 0.7131636738777161,
2948
+ "learning_rate": 4.268612627450819e-08,
2949
+ "loss": 0.0736,
2950
+ "step": 2100
2951
+ },
2952
+ {
2953
+ "epoch": 2.939944134078212,
2954
+ "grad_norm": 0.569244921207428,
2955
+ "learning_rate": 3.442229739190672e-08,
2956
+ "loss": 0.066,
2957
+ "step": 2105
2958
+ },
2959
+ {
2960
+ "epoch": 2.946927374301676,
2961
+ "grad_norm": 0.655851423740387,
2962
+ "learning_rate": 2.704576907748335e-08,
2963
+ "loss": 0.0696,
2964
+ "step": 2110
2965
+ },
2966
+ {
2967
+ "epoch": 2.9539106145251397,
2968
+ "grad_norm": 0.5713750123977661,
2969
+ "learning_rate": 2.05569786813925e-08,
2970
+ "loss": 0.0663,
2971
+ "step": 2115
2972
+ },
2973
+ {
2974
+ "epoch": 2.9608938547486034,
2975
+ "grad_norm": 0.6944213509559631,
2976
+ "learning_rate": 1.4956310920317683e-08,
2977
+ "loss": 0.0733,
2978
+ "step": 2120
2979
+ },
2980
+ {
2981
+ "epoch": 2.967877094972067,
2982
+ "grad_norm": 0.8080397248268127,
2983
+ "learning_rate": 1.024409785465641e-08,
2984
+ "loss": 0.0778,
2985
+ "step": 2125
2986
+ },
2987
+ {
2988
+ "epoch": 2.9748603351955305,
2989
+ "grad_norm": 0.729237973690033,
2990
+ "learning_rate": 6.420618868835937e-09,
2991
+ "loss": 0.0746,
2992
+ "step": 2130
2993
+ },
2994
+ {
2995
+ "epoch": 2.9818435754189943,
2996
+ "grad_norm": 0.6195108890533447,
2997
+ "learning_rate": 3.4861006547431963e-09,
2998
+ "loss": 0.0657,
2999
+ "step": 2135
3000
+ },
3001
+ {
3002
+ "epoch": 2.988826815642458,
3003
+ "grad_norm": 0.7928820848464966,
3004
+ "learning_rate": 1.4407171982905309e-09,
3005
+ "loss": 0.072,
3006
+ "step": 2140
3007
+ },
3008
+ {
3009
+ "epoch": 2.9958100558659218,
3010
+ "grad_norm": 0.6535620093345642,
3011
+ "learning_rate": 2.845897690972876e-10,
3012
+ "loss": 0.0747,
3013
+ "step": 2145
3014
+ },
3015
+ {
3016
+ "epoch": 3.0,
3017
+ "step": 2148,
3018
+ "total_flos": 2.6974388809776497e+18,
3019
+ "train_loss": 0.37994372332728776,
3020
+ "train_runtime": 1294.7645,
3021
+ "train_samples_per_second": 53.067,
3022
+ "train_steps_per_second": 1.659
3023
+ }
3024
+ ],
3025
+ "logging_steps": 5,
3026
+ "max_steps": 2148,
3027
+ "num_input_tokens_seen": 0,
3028
+ "num_train_epochs": 3,
3029
+ "save_steps": 20000,
3030
+ "stateful_callbacks": {
3031
+ "TrainerControl": {
3032
+ "args": {
3033
+ "should_epoch_stop": false,
3034
+ "should_evaluate": false,
3035
+ "should_log": false,
3036
+ "should_save": false,
3037
+ "should_training_stop": false
3038
+ },
3039
+ "attributes": {}
3040
+ }
3041
+ },
3042
+ "total_flos": 2.6974388809776497e+18,
3043
+ "train_batch_size": 2,
3044
+ "trial_name": null,
3045
+ "trial_params": null
3046
+ }
instruct/11_128_e3_3e-5/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa60ff5cf19a2545f8e6933640c3e7b50c9c6cce09cc4bf4c4049d2fc4871280
3
+ size 8273
instruct/11_128_e3_3e-5/vocab.json ADDED
The diff for this file is too large to render. See raw diff