Hugging Face's logo

RayDu0010
/
tulu10k_klora

Safetensors
Model card Files
xet
 Community
RayDu0010 commited on
Commit
05a2565
·
verified ·
1 Parent(s): 665fa65

Upload folder using huggingface_hub

Browse files
Files changed (14) hide show
  1. instruct/116_128_e3_3e-5/adapter_config.json +39 -0
  2. instruct/116_128_e3_3e-5/adapter_model.safetensors +3 -0
  3. instruct/116_128_e3_3e-5/added_tokens.json +9 -0
  4. instruct/116_128_e3_3e-5/all_results.json +9 -0
  5. instruct/116_128_e3_3e-5/chat_template.jinja +62 -0
  6. instruct/116_128_e3_3e-5/config.json +32 -0
  7. instruct/116_128_e3_3e-5/merges.txt +0 -0
  8. instruct/116_128_e3_3e-5/special_tokens_map.json +33 -0
  9. instruct/116_128_e3_3e-5/tokenizer.json +0 -0
  10. instruct/116_128_e3_3e-5/tokenizer_config.json +234 -0
  11. instruct/116_128_e3_3e-5/train_results.json +9 -0
  12. instruct/116_128_e3_3e-5/trainer_state.json +3445 -0
  13. instruct/116_128_e3_3e-5/training_args.bin +3 -0
  14. instruct/116_128_e3_3e-5/vocab.json +0 -0
instruct/116_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
+ "q_proj",
28
+ "k_proj",
29
+ "down_proj",
30
+ "v_proj",
31
+ "gate_proj",
32
+ "o_proj",
33
+ "up_proj"
34
+ ],
35
+ "task_type": "CAUSAL_LM",
36
+ "trainable_token_indices": null,
37
+ "use_dora": false,
38
+ "use_rslora": false
39
+ }
instruct/116_128_e3_3e-5/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8889a94ca52e6e4625faa6ba9d1963c5029bce33d043eb6c5057b8d3d711d683
3
+ size 791751704
instruct/116_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/116_128_e3_3e-5/all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 3.277186621681697e+18,
4
+ "train_loss": 0.3510698379803096,
5
+ "train_runtime": 1515.331,
6
+ "train_samples": 25897,
7
+ "train_samples_per_second": 51.27,
8
+ "train_steps_per_second": 1.604
9
+ }
instruct/116_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/116_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/116_128_e3_3e-5/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
instruct/116_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/116_128_e3_3e-5/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
instruct/116_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/116_128_e3_3e-5/train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 3.277186621681697e+18,
4
+ "train_loss": 0.3510698379803096,
5
+ "train_runtime": 1515.331,
6
+ "train_samples": 25897,
7
+ "train_samples_per_second": 51.27,
8
+ "train_steps_per_second": 1.604
9
+ }
instruct/116_128_e3_3e-5/trainer_state.json ADDED
@@ -0,0 +1,3445 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 2430,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.006176652254478073,
14
+ "grad_norm": 2.313079595565796,
15
+ "learning_rate": 9.836065573770493e-07,
16
+ "loss": 1.5062,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.012353304508956145,
21
+ "grad_norm": 1.9012178182601929,
22
+ "learning_rate": 2.213114754098361e-06,
23
+ "loss": 1.4976,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.01852995676343422,
28
+ "grad_norm": 1.632373332977295,
29
+ "learning_rate": 3.4426229508196724e-06,
30
+ "loss": 1.4677,
31
+ "step": 15
32
+ },
33
+ {
34
+ "epoch": 0.02470660901791229,
35
+ "grad_norm": 0.9481589198112488,
36
+ "learning_rate": 4.672131147540984e-06,
37
+ "loss": 1.4521,
38
+ "step": 20
39
+ },
40
+ {
41
+ "epoch": 0.030883261272390366,
42
+ "grad_norm": 0.5198933482170105,
43
+ "learning_rate": 5.901639344262295e-06,
44
+ "loss": 1.3998,
45
+ "step": 25
46
+ },
47
+ {
48
+ "epoch": 0.03705991352686844,
49
+ "grad_norm": 0.46687209606170654,
50
+ "learning_rate": 7.131147540983606e-06,
51
+ "loss": 1.4123,
52
+ "step": 30
53
+ },
54
+ {
55
+ "epoch": 0.04323656578134651,
56
+ "grad_norm": 0.4322935938835144,
57
+ "learning_rate": 8.360655737704917e-06,
58
+ "loss": 1.3939,
59
+ "step": 35
60
+ },
61
+ {
62
+ "epoch": 0.04941321803582458,
63
+ "grad_norm": 0.3889077603816986,
64
+ "learning_rate": 9.590163934426231e-06,
65
+ "loss": 1.3289,
66
+ "step": 40
67
+ },
68
+ {
69
+ "epoch": 0.055589870290302656,
70
+ "grad_norm": 0.4080551564693451,
71
+ "learning_rate": 1.0819672131147542e-05,
72
+ "loss": 1.3673,
73
+ "step": 45
74
+ },
75
+ {
76
+ "epoch": 0.06176652254478073,
77
+ "grad_norm": 0.3506929874420166,
78
+ "learning_rate": 1.2049180327868853e-05,
79
+ "loss": 1.3272,
80
+ "step": 50
81
+ },
82
+ {
83
+ "epoch": 0.06794317479925881,
84
+ "grad_norm": 0.4541415274143219,
85
+ "learning_rate": 1.3278688524590163e-05,
86
+ "loss": 1.326,
87
+ "step": 55
88
+ },
89
+ {
90
+ "epoch": 0.07411982705373688,
91
+ "grad_norm": 0.35612860321998596,
92
+ "learning_rate": 1.4508196721311476e-05,
93
+ "loss": 1.2876,
94
+ "step": 60
95
+ },
96
+ {
97
+ "epoch": 0.08029647930821494,
98
+ "grad_norm": 0.3479231894016266,
99
+ "learning_rate": 1.5737704918032788e-05,
100
+ "loss": 1.2836,
101
+ "step": 65
102
+ },
103
+ {
104
+ "epoch": 0.08647313156269303,
105
+ "grad_norm": 0.3883479833602905,
106
+ "learning_rate": 1.6967213114754097e-05,
107
+ "loss": 1.313,
108
+ "step": 70
109
+ },
110
+ {
111
+ "epoch": 0.0926497838171711,
112
+ "grad_norm": 0.4051356613636017,
113
+ "learning_rate": 1.819672131147541e-05,
114
+ "loss": 1.3129,
115
+ "step": 75
116
+ },
117
+ {
118
+ "epoch": 0.09882643607164916,
119
+ "grad_norm": 0.40156057476997375,
120
+ "learning_rate": 1.9426229508196722e-05,
121
+ "loss": 1.2894,
122
+ "step": 80
123
+ },
124
+ {
125
+ "epoch": 0.10500308832612724,
126
+ "grad_norm": 0.4190380275249481,
127
+ "learning_rate": 2.0655737704918034e-05,
128
+ "loss": 1.2738,
129
+ "step": 85
130
+ },
131
+ {
132
+ "epoch": 0.11117974058060531,
133
+ "grad_norm": 0.41614165902137756,
134
+ "learning_rate": 2.1885245901639347e-05,
135
+ "loss": 1.291,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.11735639283508338,
140
+ "grad_norm": 0.42249974608421326,
141
+ "learning_rate": 2.3114754098360656e-05,
142
+ "loss": 1.2429,
143
+ "step": 95
144
+ },
145
+ {
146
+ "epoch": 0.12353304508956146,
147
+ "grad_norm": 0.43746909499168396,
148
+ "learning_rate": 2.434426229508197e-05,
149
+ "loss": 1.2129,
150
+ "step": 100
151
+ },
152
+ {
153
+ "epoch": 0.12970969734403953,
154
+ "grad_norm": 0.4513344466686249,
155
+ "learning_rate": 2.5573770491803277e-05,
156
+ "loss": 1.2333,
157
+ "step": 105
158
+ },
159
+ {
160
+ "epoch": 0.13588634959851761,
161
+ "grad_norm": 0.4328799247741699,
162
+ "learning_rate": 2.680327868852459e-05,
163
+ "loss": 1.2562,
164
+ "step": 110
165
+ },
166
+ {
167
+ "epoch": 0.14206300185299567,
168
+ "grad_norm": 0.4656026065349579,
169
+ "learning_rate": 2.8032786885245902e-05,
170
+ "loss": 1.2077,
171
+ "step": 115
172
+ },
173
+ {
174
+ "epoch": 0.14823965410747375,
175
+ "grad_norm": 0.4609512984752655,
176
+ "learning_rate": 2.9262295081967215e-05,
177
+ "loss": 1.1861,
178
+ "step": 120
179
+ },
180
+ {
181
+ "epoch": 0.15441630636195183,
182
+ "grad_norm": 0.5226691961288452,
183
+ "learning_rate": 2.9999944416086244e-05,
184
+ "loss": 1.2109,
185
+ "step": 125
186
+ },
187
+ {
188
+ "epoch": 0.1605929586164299,
189
+ "grad_norm": 0.5427550077438354,
190
+ "learning_rate": 2.999931910178738e-05,
191
+ "loss": 1.1706,
192
+ "step": 130
193
+ },
194
+ {
195
+ "epoch": 0.16676961087090797,
196
+ "grad_norm": 0.5592103004455566,
197
+ "learning_rate": 2.9997999022358424e-05,
198
+ "loss": 1.1918,
199
+ "step": 135
200
+ },
201
+ {
202
+ "epoch": 0.17294626312538605,
203
+ "grad_norm": 0.5529078841209412,
204
+ "learning_rate": 2.999598423894516e-05,
205
+ "loss": 1.122,
206
+ "step": 140
207
+ },
208
+ {
209
+ "epoch": 0.1791229153798641,
210
+ "grad_norm": 0.5283639430999756,
211
+ "learning_rate": 2.9993274844871917e-05,
212
+ "loss": 1.105,
213
+ "step": 145
214
+ },
215
+ {
216
+ "epoch": 0.1852995676343422,
217
+ "grad_norm": 0.5931684970855713,
218
+ "learning_rate": 2.9989870965637215e-05,
219
+ "loss": 1.1171,
220
+ "step": 150
221
+ },
222
+ {
223
+ "epoch": 0.19147621988882027,
224
+ "grad_norm": 0.5643950700759888,
225
+ "learning_rate": 2.9985772758907992e-05,
226
+ "loss": 1.1494,
227
+ "step": 155
228
+ },
229
+ {
230
+ "epoch": 0.19765287214329832,
231
+ "grad_norm": 0.5597105622291565,
232
+ "learning_rate": 2.998098041451227e-05,
233
+ "loss": 1.1327,
234
+ "step": 160
235
+ },
236
+ {
237
+ "epoch": 0.2038295243977764,
238
+ "grad_norm": 0.5825310945510864,
239
+ "learning_rate": 2.9975494154430378e-05,
240
+ "loss": 1.111,
241
+ "step": 165
242
+ },
243
+ {
244
+ "epoch": 0.2100061766522545,
245
+ "grad_norm": 0.6192973852157593,
246
+ "learning_rate": 2.996931423278467e-05,
247
+ "loss": 1.1114,
248
+ "step": 170
249
+ },
250
+ {
251
+ "epoch": 0.21618282890673254,
252
+ "grad_norm": 0.6072234511375427,
253
+ "learning_rate": 2.9962440935827735e-05,
254
+ "loss": 1.0653,
255
+ "step": 175
256
+ },
257
+ {
258
+ "epoch": 0.22235948116121063,
259
+ "grad_norm": 0.6854657530784607,
260
+ "learning_rate": 2.995487458192917e-05,
261
+ "loss": 1.0608,
262
+ "step": 180
263
+ },
264
+ {
265
+ "epoch": 0.2285361334156887,
266
+ "grad_norm": 0.6361896395683289,
267
+ "learning_rate": 2.9946615521560805e-05,
268
+ "loss": 1.1144,
269
+ "step": 185
270
+ },
271
+ {
272
+ "epoch": 0.23471278567016676,
273
+ "grad_norm": 0.6523696780204773,
274
+ "learning_rate": 2.9937664137280478e-05,
275
+ "loss": 1.0676,
276
+ "step": 190
277
+ },
278
+ {
279
+ "epoch": 0.24088943792464484,
280
+ "grad_norm": 0.7004927396774292,
281
+ "learning_rate": 2.9928020843714323e-05,
282
+ "loss": 1.0218,
283
+ "step": 195
284
+ },
285
+ {
286
+ "epoch": 0.24706609017912293,
287
+ "grad_norm": 0.735270082950592,
288
+ "learning_rate": 2.9917686087537563e-05,
289
+ "loss": 0.9841,
290
+ "step": 200
291
+ },
292
+ {
293
+ "epoch": 0.253242742433601,
294
+ "grad_norm": 0.750301718711853,
295
+ "learning_rate": 2.9906660347453803e-05,
296
+ "loss": 1.0333,
297
+ "step": 205
298
+ },
299
+ {
300
+ "epoch": 0.25941939468807906,
301
+ "grad_norm": 0.6790189146995544,
302
+ "learning_rate": 2.9894944134172876e-05,
303
+ "loss": 1.0151,
304
+ "step": 210
305
+ },
306
+ {
307
+ "epoch": 0.2655960469425571,
308
+ "grad_norm": 0.7323490977287292,
309
+ "learning_rate": 2.988253799038718e-05,
310
+ "loss": 0.99,
311
+ "step": 215
312
+ },
313
+ {
314
+ "epoch": 0.27177269919703523,
315
+ "grad_norm": 0.6990451216697693,
316
+ "learning_rate": 2.986944249074654e-05,
317
+ "loss": 0.9789,
318
+ "step": 220
319
+ },
320
+ {
321
+ "epoch": 0.2779493514515133,
322
+ "grad_norm": 0.7893900871276855,
323
+ "learning_rate": 2.985565824183159e-05,
324
+ "loss": 0.9487,
325
+ "step": 225
326
+ },
327
+ {
328
+ "epoch": 0.28412600370599134,
329
+ "grad_norm": 0.7792460322380066,
330
+ "learning_rate": 2.9841185882125682e-05,
331
+ "loss": 0.9749,
332
+ "step": 230
333
+ },
334
+ {
335
+ "epoch": 0.29030265596046945,
336
+ "grad_norm": 0.7602144479751587,
337
+ "learning_rate": 2.9826026081985305e-05,
338
+ "loss": 0.995,
339
+ "step": 235
340
+ },
341
+ {
342
+ "epoch": 0.2964793082149475,
343
+ "grad_norm": 0.7584162950515747,
344
+ "learning_rate": 2.9810179543609032e-05,
345
+ "loss": 1.0084,
346
+ "step": 240
347
+ },
348
+ {
349
+ "epoch": 0.30265596046942556,
350
+ "grad_norm": 0.8644339442253113,
351
+ "learning_rate": 2.9793647001005002e-05,
352
+ "loss": 0.9244,
353
+ "step": 245
354
+ },
355
+ {
356
+ "epoch": 0.30883261272390367,
357
+ "grad_norm": 0.8664239645004272,
358
+ "learning_rate": 2.9776429219956917e-05,
359
+ "loss": 0.8731,
360
+ "step": 250
361
+ },
362
+ {
363
+ "epoch": 0.3150092649783817,
364
+ "grad_norm": 0.8149234056472778,
365
+ "learning_rate": 2.975852699798857e-05,
366
+ "loss": 0.909,
367
+ "step": 255
368
+ },
369
+ {
370
+ "epoch": 0.3211859172328598,
371
+ "grad_norm": 0.9656692743301392,
372
+ "learning_rate": 2.9739941164326914e-05,
373
+ "loss": 0.9098,
374
+ "step": 260
375
+ },
376
+ {
377
+ "epoch": 0.3273625694873379,
378
+ "grad_norm": 0.8673973083496094,
379
+ "learning_rate": 2.9720672579863633e-05,
380
+ "loss": 0.9046,
381
+ "step": 265
382
+ },
383
+ {
384
+ "epoch": 0.33353922174181594,
385
+ "grad_norm": 0.8363213539123535,
386
+ "learning_rate": 2.970072213711528e-05,
387
+ "loss": 0.911,
388
+ "step": 270
389
+ },
390
+ {
391
+ "epoch": 0.339715873996294,
392
+ "grad_norm": 0.949397087097168,
393
+ "learning_rate": 2.9680090760181933e-05,
394
+ "loss": 0.884,
395
+ "step": 275
396
+ },
397
+ {
398
+ "epoch": 0.3458925262507721,
399
+ "grad_norm": 0.951472282409668,
400
+ "learning_rate": 2.9658779404704387e-05,
401
+ "loss": 0.8722,
402
+ "step": 280
403
+ },
404
+ {
405
+ "epoch": 0.35206917850525016,
406
+ "grad_norm": 0.8370473384857178,
407
+ "learning_rate": 2.9636789057819903e-05,
408
+ "loss": 0.9194,
409
+ "step": 285
410
+ },
411
+ {
412
+ "epoch": 0.3582458307597282,
413
+ "grad_norm": 0.8283674120903015,
414
+ "learning_rate": 2.961412073811646e-05,
415
+ "loss": 0.9185,
416
+ "step": 290
417
+ },
418
+ {
419
+ "epoch": 0.3644224830142063,
420
+ "grad_norm": 0.9134969711303711,
421
+ "learning_rate": 2.9590775495585597e-05,
422
+ "loss": 0.8916,
423
+ "step": 295
424
+ },
425
+ {
426
+ "epoch": 0.3705991352686844,
427
+ "grad_norm": 0.9590599536895752,
428
+ "learning_rate": 2.956675441157376e-05,
429
+ "loss": 0.856,
430
+ "step": 300
431
+ },
432
+ {
433
+ "epoch": 0.37677578752316243,
434
+ "grad_norm": 0.905580997467041,
435
+ "learning_rate": 2.954205859873223e-05,
436
+ "loss": 0.8677,
437
+ "step": 305
438
+ },
439
+ {
440
+ "epoch": 0.38295243977764054,
441
+ "grad_norm": 0.966001033782959,
442
+ "learning_rate": 2.951668920096557e-05,
443
+ "loss": 0.8087,
444
+ "step": 310
445
+ },
446
+ {
447
+ "epoch": 0.3891290920321186,
448
+ "grad_norm": 0.9904844164848328,
449
+ "learning_rate": 2.9490647393378656e-05,
450
+ "loss": 0.8375,
451
+ "step": 315
452
+ },
453
+ {
454
+ "epoch": 0.39530574428659665,
455
+ "grad_norm": 1.0047897100448608,
456
+ "learning_rate": 2.9463934382222226e-05,
457
+ "loss": 0.8175,
458
+ "step": 320
459
+ },
460
+ {
461
+ "epoch": 0.40148239654107476,
462
+ "grad_norm": 0.8805506825447083,
463
+ "learning_rate": 2.943655140483703e-05,
464
+ "loss": 0.8014,
465
+ "step": 325
466
+ },
467
+ {
468
+ "epoch": 0.4076590487955528,
469
+ "grad_norm": 0.9891765117645264,
470
+ "learning_rate": 2.94084997295965e-05,
471
+ "loss": 0.8657,
472
+ "step": 330
473
+ },
474
+ {
475
+ "epoch": 0.41383570105003087,
476
+ "grad_norm": 0.9398037195205688,
477
+ "learning_rate": 2.9379780655848e-05,
478
+ "loss": 0.7971,
479
+ "step": 335
480
+ },
481
+ {
482
+ "epoch": 0.420012353304509,
483
+ "grad_norm": 1.1591811180114746,
484
+ "learning_rate": 2.9350395513852655e-05,
485
+ "loss": 0.8519,
486
+ "step": 340
487
+ },
488
+ {
489
+ "epoch": 0.42618900555898703,
490
+ "grad_norm": 0.9526787996292114,
491
+ "learning_rate": 2.9320345664723713e-05,
492
+ "loss": 0.8162,
493
+ "step": 345
494
+ },
495
+ {
496
+ "epoch": 0.4323656578134651,
497
+ "grad_norm": 1.0284078121185303,
498
+ "learning_rate": 2.928963250036351e-05,
499
+ "loss": 0.7803,
500
+ "step": 350
501
+ },
502
+ {
503
+ "epoch": 0.4385423100679432,
504
+ "grad_norm": 0.9912726283073425,
505
+ "learning_rate": 2.9258257443399007e-05,
506
+ "loss": 0.8014,
507
+ "step": 355
508
+ },
509
+ {
510
+ "epoch": 0.44471896232242125,
511
+ "grad_norm": 0.9916875958442688,
512
+ "learning_rate": 2.922622194711587e-05,
513
+ "loss": 0.7413,
514
+ "step": 360
515
+ },
516
+ {
517
+ "epoch": 0.4508956145768993,
518
+ "grad_norm": 0.9675775766372681,
519
+ "learning_rate": 2.919352749539117e-05,
520
+ "loss": 0.7425,
521
+ "step": 365
522
+ },
523
+ {
524
+ "epoch": 0.4570722668313774,
525
+ "grad_norm": 0.9666345715522766,
526
+ "learning_rate": 2.916017560262466e-05,
527
+ "loss": 0.7538,
528
+ "step": 370
529
+ },
530
+ {
531
+ "epoch": 0.46324891908585547,
532
+ "grad_norm": 1.0793135166168213,
533
+ "learning_rate": 2.91261678136686e-05,
534
+ "loss": 0.7318,
535
+ "step": 375
536
+ },
537
+ {
538
+ "epoch": 0.4694255713403335,
539
+ "grad_norm": 1.0776954889297485,
540
+ "learning_rate": 2.9091505703756224e-05,
541
+ "loss": 0.7353,
542
+ "step": 380
543
+ },
544
+ {
545
+ "epoch": 0.47560222359481164,
546
+ "grad_norm": 1.1271756887435913,
547
+ "learning_rate": 2.9056190878428766e-05,
548
+ "loss": 0.7753,
549
+ "step": 385
550
+ },
551
+ {
552
+ "epoch": 0.4817788758492897,
553
+ "grad_norm": 1.0405715703964233,
554
+ "learning_rate": 2.9020224973461098e-05,
555
+ "loss": 0.6883,
556
+ "step": 390
557
+ },
558
+ {
559
+ "epoch": 0.48795552810376774,
560
+ "grad_norm": 1.0739260911941528,
561
+ "learning_rate": 2.8983609654785948e-05,
562
+ "loss": 0.8021,
563
+ "step": 395
564
+ },
565
+ {
566
+ "epoch": 0.49413218035824585,
567
+ "grad_norm": 0.9335334897041321,
568
+ "learning_rate": 2.8946346618416742e-05,
569
+ "loss": 0.7713,
570
+ "step": 400
571
+ },
572
+ {
573
+ "epoch": 0.5003088326127239,
574
+ "grad_norm": 0.9895988702774048,
575
+ "learning_rate": 2.8908437590369056e-05,
576
+ "loss": 0.767,
577
+ "step": 405
578
+ },
579
+ {
580
+ "epoch": 0.506485484867202,
581
+ "grad_norm": 0.9870427250862122,
582
+ "learning_rate": 2.886988432658065e-05,
583
+ "loss": 0.7122,
584
+ "step": 410
585
+ },
586
+ {
587
+ "epoch": 0.5126621371216801,
588
+ "grad_norm": 1.0207875967025757,
589
+ "learning_rate": 2.8830688612830142e-05,
590
+ "loss": 0.7233,
591
+ "step": 415
592
+ },
593
+ {
594
+ "epoch": 0.5188387893761581,
595
+ "grad_norm": 1.0152679681777954,
596
+ "learning_rate": 2.8790852264654287e-05,
597
+ "loss": 0.6901,
598
+ "step": 420
599
+ },
600
+ {
601
+ "epoch": 0.5250154416306362,
602
+ "grad_norm": 1.018761396408081,
603
+ "learning_rate": 2.875037712726389e-05,
604
+ "loss": 0.6783,
605
+ "step": 425
606
+ },
607
+ {
608
+ "epoch": 0.5311920938851142,
609
+ "grad_norm": 1.1093621253967285,
610
+ "learning_rate": 2.8709265075458324e-05,
611
+ "loss": 0.6982,
612
+ "step": 430
613
+ },
614
+ {
615
+ "epoch": 0.5373687461395923,
616
+ "grad_norm": 1.1515458822250366,
617
+ "learning_rate": 2.86675180135387e-05,
618
+ "loss": 0.6494,
619
+ "step": 435
620
+ },
621
+ {
622
+ "epoch": 0.5435453983940705,
623
+ "grad_norm": 1.0824294090270996,
624
+ "learning_rate": 2.8625137875219655e-05,
625
+ "loss": 0.6928,
626
+ "step": 440
627
+ },
628
+ {
629
+ "epoch": 0.5497220506485485,
630
+ "grad_norm": 1.0219194889068604,
631
+ "learning_rate": 2.8582126623539787e-05,
632
+ "loss": 0.6306,
633
+ "step": 445
634
+ },
635
+ {
636
+ "epoch": 0.5558987029030266,
637
+ "grad_norm": 1.050694465637207,
638
+ "learning_rate": 2.8538486250770724e-05,
639
+ "loss": 0.6607,
640
+ "step": 450
641
+ },
642
+ {
643
+ "epoch": 0.5620753551575046,
644
+ "grad_norm": 1.1499159336090088,
645
+ "learning_rate": 2.849421877832484e-05,
646
+ "loss": 0.6349,
647
+ "step": 455
648
+ },
649
+ {
650
+ "epoch": 0.5682520074119827,
651
+ "grad_norm": 1.1591821908950806,
652
+ "learning_rate": 2.844932625666163e-05,
653
+ "loss": 0.625,
654
+ "step": 460
655
+ },
656
+ {
657
+ "epoch": 0.5744286596664607,
658
+ "grad_norm": 1.0648051500320435,
659
+ "learning_rate": 2.8403810765192727e-05,
660
+ "loss": 0.638,
661
+ "step": 465
662
+ },
663
+ {
664
+ "epoch": 0.5806053119209389,
665
+ "grad_norm": 1.0210459232330322,
666
+ "learning_rate": 2.835767441218559e-05,
667
+ "loss": 0.66,
668
+ "step": 470
669
+ },
670
+ {
671
+ "epoch": 0.586781964175417,
672
+ "grad_norm": 1.136823296546936,
673
+ "learning_rate": 2.831091933466584e-05,
674
+ "loss": 0.6249,
675
+ "step": 475
676
+ },
677
+ {
678
+ "epoch": 0.592958616429895,
679
+ "grad_norm": 1.1275066137313843,
680
+ "learning_rate": 2.82635476983183e-05,
681
+ "loss": 0.5986,
682
+ "step": 480
683
+ },
684
+ {
685
+ "epoch": 0.5991352686843731,
686
+ "grad_norm": 1.130084753036499,
687
+ "learning_rate": 2.821556169738663e-05,
688
+ "loss": 0.6824,
689
+ "step": 485
690
+ },
691
+ {
692
+ "epoch": 0.6053119209388511,
693
+ "grad_norm": 0.9674781560897827,
694
+ "learning_rate": 2.8166963554571748e-05,
695
+ "loss": 0.633,
696
+ "step": 490
697
+ },
698
+ {
699
+ "epoch": 0.6114885731933292,
700
+ "grad_norm": 1.0541837215423584,
701
+ "learning_rate": 2.8117755520928834e-05,
702
+ "loss": 0.6091,
703
+ "step": 495
704
+ },
705
+ {
706
+ "epoch": 0.6176652254478073,
707
+ "grad_norm": 1.0460474491119385,
708
+ "learning_rate": 2.8067939875763078e-05,
709
+ "loss": 0.5957,
710
+ "step": 500
711
+ },
712
+ {
713
+ "epoch": 0.6238418777022854,
714
+ "grad_norm": 1.228636384010315,
715
+ "learning_rate": 2.8017518926524103e-05,
716
+ "loss": 0.5957,
717
+ "step": 505
718
+ },
719
+ {
720
+ "epoch": 0.6300185299567634,
721
+ "grad_norm": 1.037697196006775,
722
+ "learning_rate": 2.7966495008699083e-05,
723
+ "loss": 0.6442,
724
+ "step": 510
725
+ },
726
+ {
727
+ "epoch": 0.6361951822112415,
728
+ "grad_norm": 0.9942713379859924,
729
+ "learning_rate": 2.7914870485704563e-05,
730
+ "loss": 0.6542,
731
+ "step": 515
732
+ },
733
+ {
734
+ "epoch": 0.6423718344657195,
735
+ "grad_norm": 1.1823099851608276,
736
+ "learning_rate": 2.7862647748776978e-05,
737
+ "loss": 0.5917,
738
+ "step": 520
739
+ },
740
+ {
741
+ "epoch": 0.6485484867201976,
742
+ "grad_norm": 0.9579377174377441,
743
+ "learning_rate": 2.780982921686191e-05,
744
+ "loss": 0.6169,
745
+ "step": 525
746
+ },
747
+ {
748
+ "epoch": 0.6547251389746758,
749
+ "grad_norm": 1.1802698373794556,
750
+ "learning_rate": 2.775641733650203e-05,
751
+ "loss": 0.5627,
752
+ "step": 530
753
+ },
754
+ {
755
+ "epoch": 0.6609017912291538,
756
+ "grad_norm": 1.1122914552688599,
757
+ "learning_rate": 2.770241458172378e-05,
758
+ "loss": 0.6033,
759
+ "step": 535
760
+ },
761
+ {
762
+ "epoch": 0.6670784434836319,
763
+ "grad_norm": 1.0568270683288574,
764
+ "learning_rate": 2.7647823453922757e-05,
765
+ "loss": 0.6007,
766
+ "step": 540
767
+ },
768
+ {
769
+ "epoch": 0.6732550957381099,
770
+ "grad_norm": 1.170803189277649,
771
+ "learning_rate": 2.7592646481747888e-05,
772
+ "loss": 0.5631,
773
+ "step": 545
774
+ },
775
+ {
776
+ "epoch": 0.679431747992588,
777
+ "grad_norm": 1.1414214372634888,
778
+ "learning_rate": 2.753688622098427e-05,
779
+ "loss": 0.612,
780
+ "step": 550
781
+ },
782
+ {
783
+ "epoch": 0.685608400247066,
784
+ "grad_norm": 1.1838423013687134,
785
+ "learning_rate": 2.7480545254434798e-05,
786
+ "loss": 0.5554,
787
+ "step": 555
788
+ },
789
+ {
790
+ "epoch": 0.6917850525015442,
791
+ "grad_norm": 1.2550277709960938,
792
+ "learning_rate": 2.7423626191800532e-05,
793
+ "loss": 0.5638,
794
+ "step": 560
795
+ },
796
+ {
797
+ "epoch": 0.6979617047560223,
798
+ "grad_norm": 1.051339030265808,
799
+ "learning_rate": 2.7366131669559816e-05,
800
+ "loss": 0.6372,
801
+ "step": 565
802
+ },
803
+ {
804
+ "epoch": 0.7041383570105003,
805
+ "grad_norm": 1.1334750652313232,
806
+ "learning_rate": 2.7308064350846148e-05,
807
+ "loss": 0.552,
808
+ "step": 570
809
+ },
810
+ {
811
+ "epoch": 0.7103150092649784,
812
+ "grad_norm": 1.1560436487197876,
813
+ "learning_rate": 2.7249426925324844e-05,
814
+ "loss": 0.578,
815
+ "step": 575
816
+ },
817
+ {
818
+ "epoch": 0.7164916615194564,
819
+ "grad_norm": 1.0834342241287231,
820
+ "learning_rate": 2.7190222109068427e-05,
821
+ "loss": 0.5631,
822
+ "step": 580
823
+ },
824
+ {
825
+ "epoch": 0.7226683137739345,
826
+ "grad_norm": 1.2761971950531006,
827
+ "learning_rate": 2.7130452644430844e-05,
828
+ "loss": 0.5216,
829
+ "step": 585
830
+ },
831
+ {
832
+ "epoch": 0.7288449660284126,
833
+ "grad_norm": 1.1653224229812622,
834
+ "learning_rate": 2.7070121299920422e-05,
835
+ "loss": 0.5325,
836
+ "step": 590
837
+ },
838
+ {
839
+ "epoch": 0.7350216182828907,
840
+ "grad_norm": 1.3431035280227661,
841
+ "learning_rate": 2.7009230870071645e-05,
842
+ "loss": 0.537,
843
+ "step": 595
844
+ },
845
+ {
846
+ "epoch": 0.7411982705373688,
847
+ "grad_norm": 1.043889045715332,
848
+ "learning_rate": 2.6947784175315694e-05,
849
+ "loss": 0.5192,
850
+ "step": 600
851
+ },
852
+ {
853
+ "epoch": 0.7473749227918468,
854
+ "grad_norm": 1.4356805086135864,
855
+ "learning_rate": 2.688578406184983e-05,
856
+ "loss": 0.536,
857
+ "step": 605
858
+ },
859
+ {
860
+ "epoch": 0.7535515750463249,
861
+ "grad_norm": 1.1674221754074097,
862
+ "learning_rate": 2.6823233401505548e-05,
863
+ "loss": 0.5092,
864
+ "step": 610
865
+ },
866
+ {
867
+ "epoch": 0.7597282273008029,
868
+ "grad_norm": 0.9921420812606812,
869
+ "learning_rate": 2.6760135091615535e-05,
870
+ "loss": 0.5251,
871
+ "step": 615
872
+ },
873
+ {
874
+ "epoch": 0.7659048795552811,
875
+ "grad_norm": 1.1863949298858643,
876
+ "learning_rate": 2.66964920548795e-05,
877
+ "loss": 0.5396,
878
+ "step": 620
879
+ },
880
+ {
881
+ "epoch": 0.7720815318097591,
882
+ "grad_norm": 1.0899137258529663,
883
+ "learning_rate": 2.663230723922877e-05,
884
+ "loss": 0.4822,
885
+ "step": 625
886
+ },
887
+ {
888
+ "epoch": 0.7782581840642372,
889
+ "grad_norm": 1.221895456314087,
890
+ "learning_rate": 2.6567583617689754e-05,
891
+ "loss": 0.4468,
892
+ "step": 630
893
+ },
894
+ {
895
+ "epoch": 0.7844348363187152,
896
+ "grad_norm": 1.2233487367630005,
897
+ "learning_rate": 2.6502324188246226e-05,
898
+ "loss": 0.506,
899
+ "step": 635
900
+ },
901
+ {
902
+ "epoch": 0.7906114885731933,
903
+ "grad_norm": 1.0487548112869263,
904
+ "learning_rate": 2.643653197370048e-05,
905
+ "loss": 0.4684,
906
+ "step": 640
907
+ },
908
+ {
909
+ "epoch": 0.7967881408276714,
910
+ "grad_norm": 1.2670767307281494,
911
+ "learning_rate": 2.6370210021533266e-05,
912
+ "loss": 0.444,
913
+ "step": 645
914
+ },
915
+ {
916
+ "epoch": 0.8029647930821495,
917
+ "grad_norm": 1.111748218536377,
918
+ "learning_rate": 2.6303361403762686e-05,
919
+ "loss": 0.479,
920
+ "step": 650
921
+ },
922
+ {
923
+ "epoch": 0.8091414453366276,
924
+ "grad_norm": 1.1121586561203003,
925
+ "learning_rate": 2.623598921680188e-05,
926
+ "loss": 0.4997,
927
+ "step": 655
928
+ },
929
+ {
930
+ "epoch": 0.8153180975911056,
931
+ "grad_norm": 1.1853457689285278,
932
+ "learning_rate": 2.6168096581315574e-05,
933
+ "loss": 0.5006,
934
+ "step": 660
935
+ },
936
+ {
937
+ "epoch": 0.8214947498455837,
938
+ "grad_norm": 1.1602599620819092,
939
+ "learning_rate": 2.609968664207558e-05,
940
+ "loss": 0.459,
941
+ "step": 665
942
+ },
943
+ {
944
+ "epoch": 0.8276714021000617,
945
+ "grad_norm": 1.2519105672836304,
946
+ "learning_rate": 2.6030762567815093e-05,
947
+ "loss": 0.4525,
948
+ "step": 670
949
+ },
950
+ {
951
+ "epoch": 0.8338480543545398,
952
+ "grad_norm": 1.0063886642456055,
953
+ "learning_rate": 2.596132755108191e-05,
954
+ "loss": 0.4482,
955
+ "step": 675
956
+ },
957
+ {
958
+ "epoch": 0.840024706609018,
959
+ "grad_norm": 1.231958031654358,
960
+ "learning_rate": 2.5891384808090608e-05,
961
+ "loss": 0.488,
962
+ "step": 680
963
+ },
964
+ {
965
+ "epoch": 0.846201358863496,
966
+ "grad_norm": 1.1607238054275513,
967
+ "learning_rate": 2.5820937578573505e-05,
968
+ "loss": 0.4253,
969
+ "step": 685
970
+ },
971
+ {
972
+ "epoch": 0.8523780111179741,
973
+ "grad_norm": 1.172845721244812,
974
+ "learning_rate": 2.5749989125630625e-05,
975
+ "loss": 0.4695,
976
+ "step": 690
977
+ },
978
+ {
979
+ "epoch": 0.8585546633724521,
980
+ "grad_norm": 1.383042573928833,
981
+ "learning_rate": 2.5678542735578553e-05,
982
+ "loss": 0.4842,
983
+ "step": 695
984
+ },
985
+ {
986
+ "epoch": 0.8647313156269302,
987
+ "grad_norm": 1.2530629634857178,
988
+ "learning_rate": 2.5606601717798212e-05,
989
+ "loss": 0.4637,
990
+ "step": 700
991
+ },
992
+ {
993
+ "epoch": 0.8709079678814082,
994
+ "grad_norm": 1.246141791343689,
995
+ "learning_rate": 2.5534169404581576e-05,
996
+ "loss": 0.4155,
997
+ "step": 705
998
+ },
999
+ {
1000
+ "epoch": 0.8770846201358864,
1001
+ "grad_norm": 1.2164368629455566,
1002
+ "learning_rate": 2.5461249150977304e-05,
1003
+ "loss": 0.4113,
1004
+ "step": 710
1005
+ },
1006
+ {
1007
+ "epoch": 0.8832612723903644,
1008
+ "grad_norm": 1.019572138786316,
1009
+ "learning_rate": 2.5387844334635356e-05,
1010
+ "loss": 0.4496,
1011
+ "step": 715
1012
+ },
1013
+ {
1014
+ "epoch": 0.8894379246448425,
1015
+ "grad_norm": 1.3600388765335083,
1016
+ "learning_rate": 2.5313958355650517e-05,
1017
+ "loss": 0.4538,
1018
+ "step": 720
1019
+ },
1020
+ {
1021
+ "epoch": 0.8956145768993206,
1022
+ "grad_norm": 1.1754273176193237,
1023
+ "learning_rate": 2.523959463640493e-05,
1024
+ "loss": 0.4151,
1025
+ "step": 725
1026
+ },
1027
+ {
1028
+ "epoch": 0.9017912291537986,
1029
+ "grad_norm": 1.2294399738311768,
1030
+ "learning_rate": 2.516475662140956e-05,
1031
+ "loss": 0.4685,
1032
+ "step": 730
1033
+ },
1034
+ {
1035
+ "epoch": 0.9079678814082767,
1036
+ "grad_norm": 1.2881512641906738,
1037
+ "learning_rate": 2.508944777714464e-05,
1038
+ "loss": 0.415,
1039
+ "step": 735
1040
+ },
1041
+ {
1042
+ "epoch": 0.9141445336627548,
1043
+ "grad_norm": 1.0499852895736694,
1044
+ "learning_rate": 2.5013671591899117e-05,
1045
+ "loss": 0.4042,
1046
+ "step": 740
1047
+ },
1048
+ {
1049
+ "epoch": 0.9203211859172329,
1050
+ "grad_norm": 1.326310157775879,
1051
+ "learning_rate": 2.4937431575609063e-05,
1052
+ "loss": 0.432,
1053
+ "step": 745
1054
+ },
1055
+ {
1056
+ "epoch": 0.9264978381717109,
1057
+ "grad_norm": 1.2161967754364014,
1058
+ "learning_rate": 2.4860731259695096e-05,
1059
+ "loss": 0.4233,
1060
+ "step": 750
1061
+ },
1062
+ {
1063
+ "epoch": 0.932674490426189,
1064
+ "grad_norm": 1.069333553314209,
1065
+ "learning_rate": 2.478357419689883e-05,
1066
+ "loss": 0.4241,
1067
+ "step": 755
1068
+ },
1069
+ {
1070
+ "epoch": 0.938851142680667,
1071
+ "grad_norm": 1.2342585325241089,
1072
+ "learning_rate": 2.4705963961118272e-05,
1073
+ "loss": 0.4614,
1074
+ "step": 760
1075
+ },
1076
+ {
1077
+ "epoch": 0.9450277949351451,
1078
+ "grad_norm": 1.2499587535858154,
1079
+ "learning_rate": 2.4627904147242315e-05,
1080
+ "loss": 0.408,
1081
+ "step": 765
1082
+ },
1083
+ {
1084
+ "epoch": 0.9512044471896233,
1085
+ "grad_norm": 1.2935307025909424,
1086
+ "learning_rate": 2.4549398370984195e-05,
1087
+ "loss": 0.3852,
1088
+ "step": 770
1089
+ },
1090
+ {
1091
+ "epoch": 0.9573810994441013,
1092
+ "grad_norm": 1.1995203495025635,
1093
+ "learning_rate": 2.447045026871405e-05,
1094
+ "loss": 0.4271,
1095
+ "step": 775
1096
+ },
1097
+ {
1098
+ "epoch": 0.9635577516985794,
1099
+ "grad_norm": 1.1074031591415405,
1100
+ "learning_rate": 2.439106349729043e-05,
1101
+ "loss": 0.3949,
1102
+ "step": 780
1103
+ },
1104
+ {
1105
+ "epoch": 0.9697344039530574,
1106
+ "grad_norm": 1.2568044662475586,
1107
+ "learning_rate": 2.431124173389097e-05,
1108
+ "loss": 0.4076,
1109
+ "step": 785
1110
+ },
1111
+ {
1112
+ "epoch": 0.9759110562075355,
1113
+ "grad_norm": 1.299391508102417,
1114
+ "learning_rate": 2.423098867584202e-05,
1115
+ "loss": 0.3736,
1116
+ "step": 790
1117
+ },
1118
+ {
1119
+ "epoch": 0.9820877084620135,
1120
+ "grad_norm": 1.326777458190918,
1121
+ "learning_rate": 2.4150308040447416e-05,
1122
+ "loss": 0.3882,
1123
+ "step": 795
1124
+ },
1125
+ {
1126
+ "epoch": 0.9882643607164917,
1127
+ "grad_norm": 1.2249014377593994,
1128
+ "learning_rate": 2.4069203564816258e-05,
1129
+ "loss": 0.4142,
1130
+ "step": 800
1131
+ },
1132
+ {
1133
+ "epoch": 0.9944410129709698,
1134
+ "grad_norm": 1.2196009159088135,
1135
+ "learning_rate": 2.3987679005689866e-05,
1136
+ "loss": 0.4034,
1137
+ "step": 805
1138
+ },
1139
+ {
1140
+ "epoch": 1.0,
1141
+ "grad_norm": 1.6548583507537842,
1142
+ "learning_rate": 2.3905738139267705e-05,
1143
+ "loss": 0.3783,
1144
+ "step": 810
1145
+ },
1146
+ {
1147
+ "epoch": 1.0061766522544782,
1148
+ "grad_norm": 1.142764925956726,
1149
+ "learning_rate": 2.3823384761032512e-05,
1150
+ "loss": 0.3092,
1151
+ "step": 815
1152
+ },
1153
+ {
1154
+ "epoch": 1.012353304508956,
1155
+ "grad_norm": 1.095353603363037,
1156
+ "learning_rate": 2.3740622685574483e-05,
1157
+ "loss": 0.3135,
1158
+ "step": 820
1159
+ },
1160
+ {
1161
+ "epoch": 1.0185299567634343,
1162
+ "grad_norm": 1.1100507974624634,
1163
+ "learning_rate": 2.3657455746414576e-05,
1164
+ "loss": 0.3037,
1165
+ "step": 825
1166
+ },
1167
+ {
1168
+ "epoch": 1.0247066090179122,
1169
+ "grad_norm": 1.1337075233459473,
1170
+ "learning_rate": 2.3573887795826942e-05,
1171
+ "loss": 0.3231,
1172
+ "step": 830
1173
+ },
1174
+ {
1175
+ "epoch": 1.0308832612723904,
1176
+ "grad_norm": 1.3290644884109497,
1177
+ "learning_rate": 2.3489922704660496e-05,
1178
+ "loss": 0.3277,
1179
+ "step": 835
1180
+ },
1181
+ {
1182
+ "epoch": 1.0370599135268685,
1183
+ "grad_norm": 0.9902516007423401,
1184
+ "learning_rate": 2.340556436215962e-05,
1185
+ "loss": 0.2995,
1186
+ "step": 840
1187
+ },
1188
+ {
1189
+ "epoch": 1.0432365657813465,
1190
+ "grad_norm": 1.1992900371551514,
1191
+ "learning_rate": 2.3320816675784e-05,
1192
+ "loss": 0.2981,
1193
+ "step": 845
1194
+ },
1195
+ {
1196
+ "epoch": 1.0494132180358247,
1197
+ "grad_norm": 1.2168498039245605,
1198
+ "learning_rate": 2.3235683571027656e-05,
1199
+ "loss": 0.2801,
1200
+ "step": 850
1201
+ },
1202
+ {
1203
+ "epoch": 1.0555898702903026,
1204
+ "grad_norm": 1.322346806526184,
1205
+ "learning_rate": 2.3150168991237088e-05,
1206
+ "loss": 0.2786,
1207
+ "step": 855
1208
+ },
1209
+ {
1210
+ "epoch": 1.0617665225447808,
1211
+ "grad_norm": 1.1575261354446411,
1212
+ "learning_rate": 2.306427689742865e-05,
1213
+ "loss": 0.304,
1214
+ "step": 860
1215
+ },
1216
+ {
1217
+ "epoch": 1.0679431747992587,
1218
+ "grad_norm": 1.1320663690567017,
1219
+ "learning_rate": 2.297801126810505e-05,
1220
+ "loss": 0.2827,
1221
+ "step": 865
1222
+ },
1223
+ {
1224
+ "epoch": 1.0741198270537369,
1225
+ "grad_norm": 1.3645118474960327,
1226
+ "learning_rate": 2.2891376099071088e-05,
1227
+ "loss": 0.2857,
1228
+ "step": 870
1229
+ },
1230
+ {
1231
+ "epoch": 1.080296479308215,
1232
+ "grad_norm": 1.1136255264282227,
1233
+ "learning_rate": 2.2804375403248547e-05,
1234
+ "loss": 0.3086,
1235
+ "step": 875
1236
+ },
1237
+ {
1238
+ "epoch": 1.086473131562693,
1239
+ "grad_norm": 1.3231781721115112,
1240
+ "learning_rate": 2.2717013210490352e-05,
1241
+ "loss": 0.3272,
1242
+ "step": 880
1243
+ },
1244
+ {
1245
+ "epoch": 1.0926497838171711,
1246
+ "grad_norm": 1.0827056169509888,
1247
+ "learning_rate": 2.2629293567393854e-05,
1248
+ "loss": 0.2764,
1249
+ "step": 885
1250
+ },
1251
+ {
1252
+ "epoch": 1.098826436071649,
1253
+ "grad_norm": 1.3882555961608887,
1254
+ "learning_rate": 2.2541220537113456e-05,
1255
+ "loss": 0.3131,
1256
+ "step": 890
1257
+ },
1258
+ {
1259
+ "epoch": 1.1050030883261273,
1260
+ "grad_norm": 1.1794378757476807,
1261
+ "learning_rate": 2.2452798199172358e-05,
1262
+ "loss": 0.2974,
1263
+ "step": 895
1264
+ },
1265
+ {
1266
+ "epoch": 1.1111797405806052,
1267
+ "grad_norm": 1.0875461101531982,
1268
+ "learning_rate": 2.2364030649273624e-05,
1269
+ "loss": 0.2789,
1270
+ "step": 900
1271
+ },
1272
+ {
1273
+ "epoch": 1.1173563928350834,
1274
+ "grad_norm": 1.2514113187789917,
1275
+ "learning_rate": 2.227492199911045e-05,
1276
+ "loss": 0.2844,
1277
+ "step": 905
1278
+ },
1279
+ {
1280
+ "epoch": 1.1235330450895615,
1281
+ "grad_norm": 1.216747522354126,
1282
+ "learning_rate": 2.2185476376175718e-05,
1283
+ "loss": 0.3025,
1284
+ "step": 910
1285
+ },
1286
+ {
1287
+ "epoch": 1.1297096973440395,
1288
+ "grad_norm": 1.1376922130584717,
1289
+ "learning_rate": 2.2095697923570835e-05,
1290
+ "loss": 0.2591,
1291
+ "step": 915
1292
+ },
1293
+ {
1294
+ "epoch": 1.1358863495985176,
1295
+ "grad_norm": 1.073193073272705,
1296
+ "learning_rate": 2.200559079981377e-05,
1297
+ "loss": 0.3088,
1298
+ "step": 920
1299
+ },
1300
+ {
1301
+ "epoch": 1.1420630018529956,
1302
+ "grad_norm": 1.0846645832061768,
1303
+ "learning_rate": 2.191515917864651e-05,
1304
+ "loss": 0.2963,
1305
+ "step": 925
1306
+ },
1307
+ {
1308
+ "epoch": 1.1482396541074738,
1309
+ "grad_norm": 1.1449332237243652,
1310
+ "learning_rate": 2.182440724884165e-05,
1311
+ "loss": 0.276,
1312
+ "step": 930
1313
+ },
1314
+ {
1315
+ "epoch": 1.154416306361952,
1316
+ "grad_norm": 1.254672646522522,
1317
+ "learning_rate": 2.1733339214008428e-05,
1318
+ "loss": 0.2863,
1319
+ "step": 935
1320
+ },
1321
+ {
1322
+ "epoch": 1.1605929586164299,
1323
+ "grad_norm": 1.1154834032058716,
1324
+ "learning_rate": 2.1641959292397998e-05,
1325
+ "loss": 0.2611,
1326
+ "step": 940
1327
+ },
1328
+ {
1329
+ "epoch": 1.166769610870908,
1330
+ "grad_norm": 1.2798595428466797,
1331
+ "learning_rate": 2.1550271716708038e-05,
1332
+ "loss": 0.2439,
1333
+ "step": 945
1334
+ },
1335
+ {
1336
+ "epoch": 1.172946263125386,
1337
+ "grad_norm": 1.140097975730896,
1338
+ "learning_rate": 2.14582807338867e-05,
1339
+ "loss": 0.2714,
1340
+ "step": 950
1341
+ },
1342
+ {
1343
+ "epoch": 1.1791229153798641,
1344
+ "grad_norm": 1.1251535415649414,
1345
+ "learning_rate": 2.1365990604935865e-05,
1346
+ "loss": 0.2893,
1347
+ "step": 955
1348
+ },
1349
+ {
1350
+ "epoch": 1.1852995676343423,
1351
+ "grad_norm": 1.3432443141937256,
1352
+ "learning_rate": 2.127340560471383e-05,
1353
+ "loss": 0.2614,
1354
+ "step": 960
1355
+ },
1356
+ {
1357
+ "epoch": 1.1914762198888202,
1358
+ "grad_norm": 1.3054141998291016,
1359
+ "learning_rate": 2.118053002173723e-05,
1360
+ "loss": 0.2744,
1361
+ "step": 965
1362
+ },
1363
+ {
1364
+ "epoch": 1.1976528721432984,
1365
+ "grad_norm": 1.1025422811508179,
1366
+ "learning_rate": 2.1087368157982456e-05,
1367
+ "loss": 0.2458,
1368
+ "step": 970
1369
+ },
1370
+ {
1371
+ "epoch": 1.2038295243977764,
1372
+ "grad_norm": 1.2638951539993286,
1373
+ "learning_rate": 2.0993924328686353e-05,
1374
+ "loss": 0.2776,
1375
+ "step": 975
1376
+ },
1377
+ {
1378
+ "epoch": 1.2100061766522545,
1379
+ "grad_norm": 1.2865864038467407,
1380
+ "learning_rate": 2.090020286214633e-05,
1381
+ "loss": 0.2822,
1382
+ "step": 980
1383
+ },
1384
+ {
1385
+ "epoch": 1.2161828289067325,
1386
+ "grad_norm": 1.1536531448364258,
1387
+ "learning_rate": 2.0806208099519916e-05,
1388
+ "loss": 0.246,
1389
+ "step": 985
1390
+ },
1391
+ {
1392
+ "epoch": 1.2223594811612106,
1393
+ "grad_norm": 1.5075135231018066,
1394
+ "learning_rate": 2.0711944394623642e-05,
1395
+ "loss": 0.2451,
1396
+ "step": 990
1397
+ },
1398
+ {
1399
+ "epoch": 1.2285361334156888,
1400
+ "grad_norm": 1.0720267295837402,
1401
+ "learning_rate": 2.0617416113731385e-05,
1402
+ "loss": 0.2261,
1403
+ "step": 995
1404
+ },
1405
+ {
1406
+ "epoch": 1.2347127856701667,
1407
+ "grad_norm": 1.0917315483093262,
1408
+ "learning_rate": 2.0522627635372126e-05,
1409
+ "loss": 0.2349,
1410
+ "step": 1000
1411
+ },
1412
+ {
1413
+ "epoch": 1.240889437924645,
1414
+ "grad_norm": 1.2396665811538696,
1415
+ "learning_rate": 2.0427583350127134e-05,
1416
+ "loss": 0.2652,
1417
+ "step": 1005
1418
+ },
1419
+ {
1420
+ "epoch": 1.2470660901791228,
1421
+ "grad_norm": 1.1459747552871704,
1422
+ "learning_rate": 2.0332287660426598e-05,
1423
+ "loss": 0.2392,
1424
+ "step": 1010
1425
+ },
1426
+ {
1427
+ "epoch": 1.253242742433601,
1428
+ "grad_norm": 1.2529373168945312,
1429
+ "learning_rate": 2.0236744980345704e-05,
1430
+ "loss": 0.2394,
1431
+ "step": 1015
1432
+ },
1433
+ {
1434
+ "epoch": 1.259419394688079,
1435
+ "grad_norm": 1.2973624467849731,
1436
+ "learning_rate": 2.014095973540017e-05,
1437
+ "loss": 0.237,
1438
+ "step": 1020
1439
+ },
1440
+ {
1441
+ "epoch": 1.2655960469425571,
1442
+ "grad_norm": 1.1186257600784302,
1443
+ "learning_rate": 2.0044936362341288e-05,
1444
+ "loss": 0.2328,
1445
+ "step": 1025
1446
+ },
1447
+ {
1448
+ "epoch": 1.2717726991970353,
1449
+ "grad_norm": 1.2702299356460571,
1450
+ "learning_rate": 1.9948679308950358e-05,
1451
+ "loss": 0.2196,
1452
+ "step": 1030
1453
+ },
1454
+ {
1455
+ "epoch": 1.2779493514515132,
1456
+ "grad_norm": 1.0467753410339355,
1457
+ "learning_rate": 1.985219303383273e-05,
1458
+ "loss": 0.2522,
1459
+ "step": 1035
1460
+ },
1461
+ {
1462
+ "epoch": 1.2841260037059914,
1463
+ "grad_norm": 1.2115217447280884,
1464
+ "learning_rate": 1.9755482006211237e-05,
1465
+ "loss": 0.2141,
1466
+ "step": 1040
1467
+ },
1468
+ {
1469
+ "epoch": 1.2903026559604696,
1470
+ "grad_norm": 1.2209748029708862,
1471
+ "learning_rate": 1.96585507057192e-05,
1472
+ "loss": 0.2498,
1473
+ "step": 1045
1474
+ },
1475
+ {
1476
+ "epoch": 1.2964793082149475,
1477
+ "grad_norm": 1.1097275018692017,
1478
+ "learning_rate": 1.9561403622192948e-05,
1479
+ "loss": 0.2542,
1480
+ "step": 1050
1481
+ },
1482
+ {
1483
+ "epoch": 1.3026559604694254,
1484
+ "grad_norm": 1.0751670598983765,
1485
+ "learning_rate": 1.9464045255463797e-05,
1486
+ "loss": 0.2252,
1487
+ "step": 1055
1488
+ },
1489
+ {
1490
+ "epoch": 1.3088326127239036,
1491
+ "grad_norm": 1.2054475545883179,
1492
+ "learning_rate": 1.936648011514969e-05,
1493
+ "loss": 0.2589,
1494
+ "step": 1060
1495
+ },
1496
+ {
1497
+ "epoch": 1.3150092649783818,
1498
+ "grad_norm": 1.1218714714050293,
1499
+ "learning_rate": 1.9268712720446255e-05,
1500
+ "loss": 0.229,
1501
+ "step": 1065
1502
+ },
1503
+ {
1504
+ "epoch": 1.3211859172328597,
1505
+ "grad_norm": 1.2714215517044067,
1506
+ "learning_rate": 1.9170747599917507e-05,
1507
+ "loss": 0.2312,
1508
+ "step": 1070
1509
+ },
1510
+ {
1511
+ "epoch": 1.3273625694873379,
1512
+ "grad_norm": 1.0570261478424072,
1513
+ "learning_rate": 1.9072589291286073e-05,
1514
+ "loss": 0.2209,
1515
+ "step": 1075
1516
+ },
1517
+ {
1518
+ "epoch": 1.333539221741816,
1519
+ "grad_norm": 1.1266241073608398,
1520
+ "learning_rate": 1.897424234122302e-05,
1521
+ "loss": 0.2411,
1522
+ "step": 1080
1523
+ },
1524
+ {
1525
+ "epoch": 1.339715873996294,
1526
+ "grad_norm": 1.0998766422271729,
1527
+ "learning_rate": 1.8875711305137222e-05,
1528
+ "loss": 0.2129,
1529
+ "step": 1085
1530
+ },
1531
+ {
1532
+ "epoch": 1.3458925262507722,
1533
+ "grad_norm": 1.1885435581207275,
1534
+ "learning_rate": 1.8777000746964397e-05,
1535
+ "loss": 0.2448,
1536
+ "step": 1090
1537
+ },
1538
+ {
1539
+ "epoch": 1.35206917850525,
1540
+ "grad_norm": 1.063388466835022,
1541
+ "learning_rate": 1.8678115238955686e-05,
1542
+ "loss": 0.2493,
1543
+ "step": 1095
1544
+ },
1545
+ {
1546
+ "epoch": 1.3582458307597283,
1547
+ "grad_norm": 1.15679931640625,
1548
+ "learning_rate": 1.8579059361465863e-05,
1549
+ "loss": 0.2073,
1550
+ "step": 1100
1551
+ },
1552
+ {
1553
+ "epoch": 1.3644224830142062,
1554
+ "grad_norm": 1.1405763626098633,
1555
+ "learning_rate": 1.8479837702741166e-05,
1556
+ "loss": 0.2038,
1557
+ "step": 1105
1558
+ },
1559
+ {
1560
+ "epoch": 1.3705991352686844,
1561
+ "grad_norm": 1.1770544052124023,
1562
+ "learning_rate": 1.8380454858706798e-05,
1563
+ "loss": 0.2172,
1564
+ "step": 1110
1565
+ },
1566
+ {
1567
+ "epoch": 1.3767757875231625,
1568
+ "grad_norm": 1.3772873878479004,
1569
+ "learning_rate": 1.8280915432754033e-05,
1570
+ "loss": 0.2233,
1571
+ "step": 1115
1572
+ },
1573
+ {
1574
+ "epoch": 1.3829524397776405,
1575
+ "grad_norm": 1.0607002973556519,
1576
+ "learning_rate": 1.8181224035526968e-05,
1577
+ "loss": 0.2283,
1578
+ "step": 1120
1579
+ },
1580
+ {
1581
+ "epoch": 1.3891290920321187,
1582
+ "grad_norm": 1.2292882204055786,
1583
+ "learning_rate": 1.8081385284708995e-05,
1584
+ "loss": 0.2142,
1585
+ "step": 1125
1586
+ },
1587
+ {
1588
+ "epoch": 1.3953057442865966,
1589
+ "grad_norm": 1.3323862552642822,
1590
+ "learning_rate": 1.798140380480888e-05,
1591
+ "loss": 0.2224,
1592
+ "step": 1130
1593
+ },
1594
+ {
1595
+ "epoch": 1.4014823965410748,
1596
+ "grad_norm": 1.0667893886566162,
1597
+ "learning_rate": 1.7881284226946558e-05,
1598
+ "loss": 0.1744,
1599
+ "step": 1135
1600
+ },
1601
+ {
1602
+ "epoch": 1.4076590487955527,
1603
+ "grad_norm": 1.2054765224456787,
1604
+ "learning_rate": 1.778103118863866e-05,
1605
+ "loss": 0.2292,
1606
+ "step": 1140
1607
+ },
1608
+ {
1609
+ "epoch": 1.4138357010500309,
1610
+ "grad_norm": 0.9714310765266418,
1611
+ "learning_rate": 1.768064933358365e-05,
1612
+ "loss": 0.1912,
1613
+ "step": 1145
1614
+ },
1615
+ {
1616
+ "epoch": 1.420012353304509,
1617
+ "grad_norm": 1.1717028617858887,
1618
+ "learning_rate": 1.7580143311446766e-05,
1619
+ "loss": 0.2047,
1620
+ "step": 1150
1621
+ },
1622
+ {
1623
+ "epoch": 1.426189005558987,
1624
+ "grad_norm": 1.155126929283142,
1625
+ "learning_rate": 1.7479517777644632e-05,
1626
+ "loss": 0.2364,
1627
+ "step": 1155
1628
+ },
1629
+ {
1630
+ "epoch": 1.4323656578134651,
1631
+ "grad_norm": 1.0830705165863037,
1632
+ "learning_rate": 1.7378777393129635e-05,
1633
+ "loss": 0.1898,
1634
+ "step": 1160
1635
+ },
1636
+ {
1637
+ "epoch": 1.4385423100679433,
1638
+ "grad_norm": 1.0223325490951538,
1639
+ "learning_rate": 1.7277926824174012e-05,
1640
+ "loss": 0.2116,
1641
+ "step": 1165
1642
+ },
1643
+ {
1644
+ "epoch": 1.4447189623224213,
1645
+ "grad_norm": 1.0795882940292358,
1646
+ "learning_rate": 1.7176970742153717e-05,
1647
+ "loss": 0.1883,
1648
+ "step": 1170
1649
+ },
1650
+ {
1651
+ "epoch": 1.4508956145768992,
1652
+ "grad_norm": 1.0658038854599,
1653
+ "learning_rate": 1.7075913823332067e-05,
1654
+ "loss": 0.1885,
1655
+ "step": 1175
1656
+ },
1657
+ {
1658
+ "epoch": 1.4570722668313774,
1659
+ "grad_norm": 1.0698961019515991,
1660
+ "learning_rate": 1.6974760748643084e-05,
1661
+ "loss": 0.225,
1662
+ "step": 1180
1663
+ },
1664
+ {
1665
+ "epoch": 1.4632489190858555,
1666
+ "grad_norm": 1.1288405656814575,
1667
+ "learning_rate": 1.6873516203474736e-05,
1668
+ "loss": 0.1971,
1669
+ "step": 1185
1670
+ },
1671
+ {
1672
+ "epoch": 1.4694255713403335,
1673
+ "grad_norm": 1.1523617506027222,
1674
+ "learning_rate": 1.677218487745187e-05,
1675
+ "loss": 0.1634,
1676
+ "step": 1190
1677
+ },
1678
+ {
1679
+ "epoch": 1.4756022235948116,
1680
+ "grad_norm": 1.3232675790786743,
1681
+ "learning_rate": 1.6670771464219002e-05,
1682
+ "loss": 0.1879,
1683
+ "step": 1195
1684
+ },
1685
+ {
1686
+ "epoch": 1.4817788758492898,
1687
+ "grad_norm": 1.0664860010147095,
1688
+ "learning_rate": 1.6569280661222924e-05,
1689
+ "loss": 0.1669,
1690
+ "step": 1200
1691
+ },
1692
+ {
1693
+ "epoch": 1.4879555281037677,
1694
+ "grad_norm": 1.6697794198989868,
1695
+ "learning_rate": 1.646771716949509e-05,
1696
+ "loss": 0.2016,
1697
+ "step": 1205
1698
+ },
1699
+ {
1700
+ "epoch": 1.494132180358246,
1701
+ "grad_norm": 1.1751537322998047,
1702
+ "learning_rate": 1.636608569343387e-05,
1703
+ "loss": 0.1894,
1704
+ "step": 1210
1705
+ },
1706
+ {
1707
+ "epoch": 1.5003088326127239,
1708
+ "grad_norm": 1.1710859537124634,
1709
+ "learning_rate": 1.626439094058669e-05,
1710
+ "loss": 0.2012,
1711
+ "step": 1215
1712
+ },
1713
+ {
1714
+ "epoch": 1.506485484867202,
1715
+ "grad_norm": 1.0292590856552124,
1716
+ "learning_rate": 1.6162637621431906e-05,
1717
+ "loss": 0.1671,
1718
+ "step": 1220
1719
+ },
1720
+ {
1721
+ "epoch": 1.51266213712168,
1722
+ "grad_norm": 1.180564522743225,
1723
+ "learning_rate": 1.606083044916067e-05,
1724
+ "loss": 0.1906,
1725
+ "step": 1225
1726
+ },
1727
+ {
1728
+ "epoch": 1.5188387893761581,
1729
+ "grad_norm": 0.9680458903312683,
1730
+ "learning_rate": 1.595897413945859e-05,
1731
+ "loss": 0.1721,
1732
+ "step": 1230
1733
+ },
1734
+ {
1735
+ "epoch": 1.5250154416306363,
1736
+ "grad_norm": 1.223307490348816,
1737
+ "learning_rate": 1.5857073410287317e-05,
1738
+ "loss": 0.2026,
1739
+ "step": 1235
1740
+ },
1741
+ {
1742
+ "epoch": 1.5311920938851142,
1743
+ "grad_norm": 1.1718299388885498,
1744
+ "learning_rate": 1.5755132981666004e-05,
1745
+ "loss": 0.1771,
1746
+ "step": 1240
1747
+ },
1748
+ {
1749
+ "epoch": 1.5373687461395922,
1750
+ "grad_norm": 1.1107652187347412,
1751
+ "learning_rate": 1.5653157575452656e-05,
1752
+ "loss": 0.166,
1753
+ "step": 1245
1754
+ },
1755
+ {
1756
+ "epoch": 1.5435453983940706,
1757
+ "grad_norm": 1.0210204124450684,
1758
+ "learning_rate": 1.555115191512546e-05,
1759
+ "loss": 0.1912,
1760
+ "step": 1250
1761
+ },
1762
+ {
1763
+ "epoch": 1.5497220506485485,
1764
+ "grad_norm": 1.0872509479522705,
1765
+ "learning_rate": 1.5449120725563932e-05,
1766
+ "loss": 0.1688,
1767
+ "step": 1255
1768
+ },
1769
+ {
1770
+ "epoch": 1.5558987029030265,
1771
+ "grad_norm": 1.1102087497711182,
1772
+ "learning_rate": 1.5347068732830134e-05,
1773
+ "loss": 0.1828,
1774
+ "step": 1260
1775
+ },
1776
+ {
1777
+ "epoch": 1.5620753551575046,
1778
+ "grad_norm": 0.9405921697616577,
1779
+ "learning_rate": 1.5245000663949699e-05,
1780
+ "loss": 0.1779,
1781
+ "step": 1265
1782
+ },
1783
+ {
1784
+ "epoch": 1.5682520074119828,
1785
+ "grad_norm": 1.143991231918335,
1786
+ "learning_rate": 1.5142921246692909e-05,
1787
+ "loss": 0.1576,
1788
+ "step": 1270
1789
+ },
1790
+ {
1791
+ "epoch": 1.5744286596664607,
1792
+ "grad_norm": 1.153887391090393,
1793
+ "learning_rate": 1.5040835209355715e-05,
1794
+ "loss": 0.1636,
1795
+ "step": 1275
1796
+ },
1797
+ {
1798
+ "epoch": 1.580605311920939,
1799
+ "grad_norm": 0.9921665191650391,
1800
+ "learning_rate": 1.4938747280540692e-05,
1801
+ "loss": 0.1496,
1802
+ "step": 1280
1803
+ },
1804
+ {
1805
+ "epoch": 1.586781964175417,
1806
+ "grad_norm": 1.0411317348480225,
1807
+ "learning_rate": 1.4836662188938028e-05,
1808
+ "loss": 0.1654,
1809
+ "step": 1285
1810
+ },
1811
+ {
1812
+ "epoch": 1.592958616429895,
1813
+ "grad_norm": 1.0908194780349731,
1814
+ "learning_rate": 1.4734584663106506e-05,
1815
+ "loss": 0.1625,
1816
+ "step": 1290
1817
+ },
1818
+ {
1819
+ "epoch": 1.599135268684373,
1820
+ "grad_norm": 0.9990919828414917,
1821
+ "learning_rate": 1.4632519431254446e-05,
1822
+ "loss": 0.1738,
1823
+ "step": 1295
1824
+ },
1825
+ {
1826
+ "epoch": 1.6053119209388511,
1827
+ "grad_norm": 1.3530173301696777,
1828
+ "learning_rate": 1.453047122102073e-05,
1829
+ "loss": 0.1876,
1830
+ "step": 1300
1831
+ },
1832
+ {
1833
+ "epoch": 1.6114885731933293,
1834
+ "grad_norm": 1.0382442474365234,
1835
+ "learning_rate": 1.44284447592558e-05,
1836
+ "loss": 0.1726,
1837
+ "step": 1305
1838
+ },
1839
+ {
1840
+ "epoch": 1.6176652254478072,
1841
+ "grad_norm": 1.1343178749084473,
1842
+ "learning_rate": 1.4326444771802699e-05,
1843
+ "loss": 0.1816,
1844
+ "step": 1310
1845
+ },
1846
+ {
1847
+ "epoch": 1.6238418777022854,
1848
+ "grad_norm": 1.2515641450881958,
1849
+ "learning_rate": 1.4224475983278204e-05,
1850
+ "loss": 0.19,
1851
+ "step": 1315
1852
+ },
1853
+ {
1854
+ "epoch": 1.6300185299567636,
1855
+ "grad_norm": 0.9420174956321716,
1856
+ "learning_rate": 1.412254311685395e-05,
1857
+ "loss": 0.1656,
1858
+ "step": 1320
1859
+ },
1860
+ {
1861
+ "epoch": 1.6361951822112415,
1862
+ "grad_norm": 1.0993460416793823,
1863
+ "learning_rate": 1.4020650894037673e-05,
1864
+ "loss": 0.1616,
1865
+ "step": 1325
1866
+ },
1867
+ {
1868
+ "epoch": 1.6423718344657194,
1869
+ "grad_norm": 1.2704250812530518,
1870
+ "learning_rate": 1.3918804034454521e-05,
1871
+ "loss": 0.1597,
1872
+ "step": 1330
1873
+ },
1874
+ {
1875
+ "epoch": 1.6485484867201976,
1876
+ "grad_norm": 1.0425715446472168,
1877
+ "learning_rate": 1.3817007255628402e-05,
1878
+ "loss": 0.1507,
1879
+ "step": 1335
1880
+ },
1881
+ {
1882
+ "epoch": 1.6547251389746758,
1883
+ "grad_norm": 1.062270998954773,
1884
+ "learning_rate": 1.3715265272763515e-05,
1885
+ "loss": 0.1302,
1886
+ "step": 1340
1887
+ },
1888
+ {
1889
+ "epoch": 1.6609017912291537,
1890
+ "grad_norm": 0.9804355502128601,
1891
+ "learning_rate": 1.3613582798525903e-05,
1892
+ "loss": 0.1396,
1893
+ "step": 1345
1894
+ },
1895
+ {
1896
+ "epoch": 1.6670784434836319,
1897
+ "grad_norm": 1.2854338884353638,
1898
+ "learning_rate": 1.3511964542825206e-05,
1899
+ "loss": 0.165,
1900
+ "step": 1350
1901
+ },
1902
+ {
1903
+ "epoch": 1.67325509573811,
1904
+ "grad_norm": 1.1114675998687744,
1905
+ "learning_rate": 1.341041521259645e-05,
1906
+ "loss": 0.161,
1907
+ "step": 1355
1908
+ },
1909
+ {
1910
+ "epoch": 1.679431747992588,
1911
+ "grad_norm": 1.122304916381836,
1912
+ "learning_rate": 1.330893951158207e-05,
1913
+ "loss": 0.1573,
1914
+ "step": 1360
1915
+ },
1916
+ {
1917
+ "epoch": 1.685608400247066,
1918
+ "grad_norm": 1.0995293855667114,
1919
+ "learning_rate": 1.3207542140114007e-05,
1920
+ "loss": 0.1509,
1921
+ "step": 1365
1922
+ },
1923
+ {
1924
+ "epoch": 1.6917850525015443,
1925
+ "grad_norm": 1.1364750862121582,
1926
+ "learning_rate": 1.310622779489598e-05,
1927
+ "loss": 0.1556,
1928
+ "step": 1370
1929
+ },
1930
+ {
1931
+ "epoch": 1.6979617047560223,
1932
+ "grad_norm": 1.0758750438690186,
1933
+ "learning_rate": 1.3005001168785973e-05,
1934
+ "loss": 0.1471,
1935
+ "step": 1375
1936
+ },
1937
+ {
1938
+ "epoch": 1.7041383570105002,
1939
+ "grad_norm": 1.0333322286605835,
1940
+ "learning_rate": 1.290386695057882e-05,
1941
+ "loss": 0.1292,
1942
+ "step": 1380
1943
+ },
1944
+ {
1945
+ "epoch": 1.7103150092649784,
1946
+ "grad_norm": 0.9372277855873108,
1947
+ "learning_rate": 1.2802829824789054e-05,
1948
+ "loss": 0.1418,
1949
+ "step": 1385
1950
+ },
1951
+ {
1952
+ "epoch": 1.7164916615194565,
1953
+ "grad_norm": 1.0819621086120605,
1954
+ "learning_rate": 1.2701894471433913e-05,
1955
+ "loss": 0.1424,
1956
+ "step": 1390
1957
+ },
1958
+ {
1959
+ "epoch": 1.7226683137739345,
1960
+ "grad_norm": 1.008872151374817,
1961
+ "learning_rate": 1.2601065565816541e-05,
1962
+ "loss": 0.1359,
1963
+ "step": 1395
1964
+ },
1965
+ {
1966
+ "epoch": 1.7288449660284126,
1967
+ "grad_norm": 1.232922911643982,
1968
+ "learning_rate": 1.2500347778309463e-05,
1969
+ "loss": 0.1348,
1970
+ "step": 1400
1971
+ },
1972
+ {
1973
+ "epoch": 1.7350216182828908,
1974
+ "grad_norm": 0.9108769297599792,
1975
+ "learning_rate": 1.2399745774138226e-05,
1976
+ "loss": 0.1244,
1977
+ "step": 1405
1978
+ },
1979
+ {
1980
+ "epoch": 1.7411982705373688,
1981
+ "grad_norm": 1.0652124881744385,
1982
+ "learning_rate": 1.2299264213165329e-05,
1983
+ "loss": 0.1347,
1984
+ "step": 1410
1985
+ },
1986
+ {
1987
+ "epoch": 1.7473749227918467,
1988
+ "grad_norm": 1.10073721408844,
1989
+ "learning_rate": 1.2198907749674355e-05,
1990
+ "loss": 0.1304,
1991
+ "step": 1415
1992
+ },
1993
+ {
1994
+ "epoch": 1.7535515750463249,
1995
+ "grad_norm": 1.0913841724395752,
1996
+ "learning_rate": 1.2098681032154417e-05,
1997
+ "loss": 0.1418,
1998
+ "step": 1420
1999
+ },
2000
+ {
2001
+ "epoch": 1.759728227300803,
2002
+ "grad_norm": 1.0939198732376099,
2003
+ "learning_rate": 1.1998588703084815e-05,
2004
+ "loss": 0.1259,
2005
+ "step": 1425
2006
+ },
2007
+ {
2008
+ "epoch": 1.765904879555281,
2009
+ "grad_norm": 1.1844559907913208,
2010
+ "learning_rate": 1.1898635398720003e-05,
2011
+ "loss": 0.1453,
2012
+ "step": 1430
2013
+ },
2014
+ {
2015
+ "epoch": 1.7720815318097591,
2016
+ "grad_norm": 0.8979021906852722,
2017
+ "learning_rate": 1.1798825748874837e-05,
2018
+ "loss": 0.1328,
2019
+ "step": 1435
2020
+ },
2021
+ {
2022
+ "epoch": 1.7782581840642373,
2023
+ "grad_norm": 0.9380536675453186,
2024
+ "learning_rate": 1.1699164376710146e-05,
2025
+ "loss": 0.1317,
2026
+ "step": 1440
2027
+ },
2028
+ {
2029
+ "epoch": 1.7844348363187152,
2030
+ "grad_norm": 1.1315381526947021,
2031
+ "learning_rate": 1.1599655898518544e-05,
2032
+ "loss": 0.1289,
2033
+ "step": 1445
2034
+ },
2035
+ {
2036
+ "epoch": 1.7906114885731932,
2037
+ "grad_norm": 1.1216084957122803,
2038
+ "learning_rate": 1.1500304923510658e-05,
2039
+ "loss": 0.1342,
2040
+ "step": 1450
2041
+ },
2042
+ {
2043
+ "epoch": 1.7967881408276714,
2044
+ "grad_norm": 1.110984206199646,
2045
+ "learning_rate": 1.1401116053601584e-05,
2046
+ "loss": 0.1354,
2047
+ "step": 1455
2048
+ },
2049
+ {
2050
+ "epoch": 1.8029647930821495,
2051
+ "grad_norm": 0.9963058233261108,
2052
+ "learning_rate": 1.1302093883197748e-05,
2053
+ "loss": 0.1266,
2054
+ "step": 1460
2055
+ },
2056
+ {
2057
+ "epoch": 1.8091414453366275,
2058
+ "grad_norm": 1.0252258777618408,
2059
+ "learning_rate": 1.1203242998984098e-05,
2060
+ "loss": 0.1263,
2061
+ "step": 1465
2062
+ },
2063
+ {
2064
+ "epoch": 1.8153180975911056,
2065
+ "grad_norm": 1.2458548545837402,
2066
+ "learning_rate": 1.1104567979711636e-05,
2067
+ "loss": 0.1484,
2068
+ "step": 1470
2069
+ },
2070
+ {
2071
+ "epoch": 1.8214947498455838,
2072
+ "grad_norm": 0.91730797290802,
2073
+ "learning_rate": 1.1006073395985345e-05,
2074
+ "loss": 0.1239,
2075
+ "step": 1475
2076
+ },
2077
+ {
2078
+ "epoch": 1.8276714021000617,
2079
+ "grad_norm": 1.2536972761154175,
2080
+ "learning_rate": 1.0907763810052478e-05,
2081
+ "loss": 0.1186,
2082
+ "step": 1480
2083
+ },
2084
+ {
2085
+ "epoch": 1.8338480543545397,
2086
+ "grad_norm": 0.8824079036712646,
2087
+ "learning_rate": 1.0809643775591215e-05,
2088
+ "loss": 0.1153,
2089
+ "step": 1485
2090
+ },
2091
+ {
2092
+ "epoch": 1.840024706609018,
2093
+ "grad_norm": 1.0572572946548462,
2094
+ "learning_rate": 1.0711717837499775e-05,
2095
+ "loss": 0.1391,
2096
+ "step": 1490
2097
+ },
2098
+ {
2099
+ "epoch": 1.846201358863496,
2100
+ "grad_norm": 1.0227898359298706,
2101
+ "learning_rate": 1.0613990531685854e-05,
2102
+ "loss": 0.1222,
2103
+ "step": 1495
2104
+ },
2105
+ {
2106
+ "epoch": 1.852378011117974,
2107
+ "grad_norm": 0.9624846577644348,
2108
+ "learning_rate": 1.0516466384856564e-05,
2109
+ "loss": 0.1139,
2110
+ "step": 1500
2111
+ },
2112
+ {
2113
+ "epoch": 1.8585546633724521,
2114
+ "grad_norm": 0.947074830532074,
2115
+ "learning_rate": 1.0419149914308723e-05,
2116
+ "loss": 0.119,
2117
+ "step": 1505
2118
+ },
2119
+ {
2120
+ "epoch": 1.8647313156269303,
2121
+ "grad_norm": 0.963910698890686,
2122
+ "learning_rate": 1.0322045627719643e-05,
2123
+ "loss": 0.1249,
2124
+ "step": 1510
2125
+ },
2126
+ {
2127
+ "epoch": 1.8709079678814082,
2128
+ "grad_norm": 0.9410413503646851,
2129
+ "learning_rate": 1.0225158022938318e-05,
2130
+ "loss": 0.1296,
2131
+ "step": 1515
2132
+ },
2133
+ {
2134
+ "epoch": 1.8770846201358864,
2135
+ "grad_norm": 1.0444504022598267,
2136
+ "learning_rate": 1.0128491587777082e-05,
2137
+ "loss": 0.1278,
2138
+ "step": 1520
2139
+ },
2140
+ {
2141
+ "epoch": 1.8832612723903646,
2142
+ "grad_norm": 0.9436255097389221,
2143
+ "learning_rate": 1.003205079980375e-05,
2144
+ "loss": 0.1312,
2145
+ "step": 1525
2146
+ },
2147
+ {
2148
+ "epoch": 1.8894379246448425,
2149
+ "grad_norm": 1.0049701929092407,
2150
+ "learning_rate": 9.9358401261342e-06,
2151
+ "loss": 0.117,
2152
+ "step": 1530
2153
+ },
2154
+ {
2155
+ "epoch": 1.8956145768993204,
2156
+ "grad_norm": 0.8774596452713013,
2157
+ "learning_rate": 9.83986402322548e-06,
2158
+ "loss": 0.1152,
2159
+ "step": 1535
2160
+ },
2161
+ {
2162
+ "epoch": 1.9017912291537986,
2163
+ "grad_norm": 1.0037027597427368,
2164
+ "learning_rate": 9.744126936669352e-06,
2165
+ "loss": 0.1167,
2166
+ "step": 1540
2167
+ },
2168
+ {
2169
+ "epoch": 1.9079678814082768,
2170
+ "grad_norm": 0.8806500434875488,
2171
+ "learning_rate": 9.648633300986418e-06,
2172
+ "loss": 0.1201,
2173
+ "step": 1545
2174
+ },
2175
+ {
2176
+ "epoch": 1.9141445336627547,
2177
+ "grad_norm": 0.9353247284889221,
2178
+ "learning_rate": 9.553387539420678e-06,
2179
+ "loss": 0.1087,
2180
+ "step": 1550
2181
+ },
2182
+ {
2183
+ "epoch": 1.9203211859172329,
2184
+ "grad_norm": 0.9986940026283264,
2185
+ "learning_rate": 9.458394063734651e-06,
2186
+ "loss": 0.1307,
2187
+ "step": 1555
2188
+ },
2189
+ {
2190
+ "epoch": 1.926497838171711,
2191
+ "grad_norm": 0.9718974232673645,
2192
+ "learning_rate": 9.36365727400504e-06,
2193
+ "loss": 0.1194,
2194
+ "step": 1560
2195
+ },
2196
+ {
2197
+ "epoch": 1.932674490426189,
2198
+ "grad_norm": 0.929785430431366,
2199
+ "learning_rate": 9.26918155841891e-06,
2200
+ "loss": 0.1206,
2201
+ "step": 1565
2202
+ },
2203
+ {
2204
+ "epoch": 1.938851142680667,
2205
+ "grad_norm": 0.944404125213623,
2206
+ "learning_rate": 9.174971293070417e-06,
2207
+ "loss": 0.1125,
2208
+ "step": 1570
2209
+ },
2210
+ {
2211
+ "epoch": 1.945027794935145,
2212
+ "grad_norm": 0.8229379057884216,
2213
+ "learning_rate": 9.08103084175815e-06,
2214
+ "loss": 0.1169,
2215
+ "step": 1575
2216
+ },
2217
+ {
2218
+ "epoch": 1.9512044471896233,
2219
+ "grad_norm": 1.004645824432373,
2220
+ "learning_rate": 8.987364555782945e-06,
2221
+ "loss": 0.1172,
2222
+ "step": 1580
2223
+ },
2224
+ {
2225
+ "epoch": 1.9573810994441012,
2226
+ "grad_norm": 0.9644873738288879,
2227
+ "learning_rate": 8.893976773746363e-06,
2228
+ "loss": 0.1053,
2229
+ "step": 1585
2230
+ },
2231
+ {
2232
+ "epoch": 1.9635577516985794,
2233
+ "grad_norm": 0.9268193244934082,
2234
+ "learning_rate": 8.800871821349739e-06,
2235
+ "loss": 0.1085,
2236
+ "step": 1590
2237
+ },
2238
+ {
2239
+ "epoch": 1.9697344039530575,
2240
+ "grad_norm": 0.9020305871963501,
2241
+ "learning_rate": 8.708054011193794e-06,
2242
+ "loss": 0.0929,
2243
+ "step": 1595
2244
+ },
2245
+ {
2246
+ "epoch": 1.9759110562075355,
2247
+ "grad_norm": 1.1831473112106323,
2248
+ "learning_rate": 8.61552764257887e-06,
2249
+ "loss": 0.109,
2250
+ "step": 1600
2251
+ },
2252
+ {
2253
+ "epoch": 1.9820877084620134,
2254
+ "grad_norm": 0.8018283247947693,
2255
+ "learning_rate": 8.52329700130583e-06,
2256
+ "loss": 0.1034,
2257
+ "step": 1605
2258
+ },
2259
+ {
2260
+ "epoch": 1.9882643607164918,
2261
+ "grad_norm": 0.8712825775146484,
2262
+ "learning_rate": 8.431366359477499e-06,
2263
+ "loss": 0.103,
2264
+ "step": 1610
2265
+ },
2266
+ {
2267
+ "epoch": 1.9944410129709698,
2268
+ "grad_norm": 0.8009131550788879,
2269
+ "learning_rate": 8.339739975300779e-06,
2270
+ "loss": 0.1014,
2271
+ "step": 1615
2272
+ },
2273
+ {
2274
+ "epoch": 2.0,
2275
+ "grad_norm": 1.4263570308685303,
2276
+ "learning_rate": 8.248422092889449e-06,
2277
+ "loss": 0.1125,
2278
+ "step": 1620
2279
+ },
2280
+ {
2281
+ "epoch": 2.006176652254478,
2282
+ "grad_norm": 0.8275551795959473,
2283
+ "learning_rate": 8.15741694206754e-06,
2284
+ "loss": 0.0764,
2285
+ "step": 1625
2286
+ },
2287
+ {
2288
+ "epoch": 2.0123533045089563,
2289
+ "grad_norm": 0.965294361114502,
2290
+ "learning_rate": 8.06672873817343e-06,
2291
+ "loss": 0.0955,
2292
+ "step": 1630
2293
+ },
2294
+ {
2295
+ "epoch": 2.0185299567634343,
2296
+ "grad_norm": 0.8140105605125427,
2297
+ "learning_rate": 7.976361681864586e-06,
2298
+ "loss": 0.0905,
2299
+ "step": 1635
2300
+ },
2301
+ {
2302
+ "epoch": 2.024706609017912,
2303
+ "grad_norm": 0.8345964550971985,
2304
+ "learning_rate": 7.886319958922992e-06,
2305
+ "loss": 0.0853,
2306
+ "step": 1640
2307
+ },
2308
+ {
2309
+ "epoch": 2.03088326127239,
2310
+ "grad_norm": 0.8345704674720764,
2311
+ "learning_rate": 7.796607740061265e-06,
2312
+ "loss": 0.0852,
2313
+ "step": 1645
2314
+ },
2315
+ {
2316
+ "epoch": 2.0370599135268685,
2317
+ "grad_norm": 0.7206991910934448,
2318
+ "learning_rate": 7.707229180729449e-06,
2319
+ "loss": 0.0839,
2320
+ "step": 1650
2321
+ },
2322
+ {
2323
+ "epoch": 2.0432365657813465,
2324
+ "grad_norm": 0.7540867924690247,
2325
+ "learning_rate": 7.618188420922568e-06,
2326
+ "loss": 0.0848,
2327
+ "step": 1655
2328
+ },
2329
+ {
2330
+ "epoch": 2.0494132180358244,
2331
+ "grad_norm": 0.9760220050811768,
2332
+ "learning_rate": 7.52948958498885e-06,
2333
+ "loss": 0.0861,
2334
+ "step": 1660
2335
+ },
2336
+ {
2337
+ "epoch": 2.055589870290303,
2338
+ "grad_norm": 0.729282796382904,
2339
+ "learning_rate": 7.441136781438677e-06,
2340
+ "loss": 0.0937,
2341
+ "step": 1665
2342
+ },
2343
+ {
2344
+ "epoch": 2.0617665225447808,
2345
+ "grad_norm": 0.9330270290374756,
2346
+ "learning_rate": 7.353134102754289e-06,
2347
+ "loss": 0.0841,
2348
+ "step": 1670
2349
+ },
2350
+ {
2351
+ "epoch": 2.0679431747992587,
2352
+ "grad_norm": 0.8869553208351135,
2353
+ "learning_rate": 7.265485625200221e-06,
2354
+ "loss": 0.0906,
2355
+ "step": 1675
2356
+ },
2357
+ {
2358
+ "epoch": 2.074119827053737,
2359
+ "grad_norm": 0.7082441449165344,
2360
+ "learning_rate": 7.1781954086344976e-06,
2361
+ "loss": 0.0807,
2362
+ "step": 1680
2363
+ },
2364
+ {
2365
+ "epoch": 2.080296479308215,
2366
+ "grad_norm": 0.981907844543457,
2367
+ "learning_rate": 7.091267496320553e-06,
2368
+ "loss": 0.0866,
2369
+ "step": 1685
2370
+ },
2371
+ {
2372
+ "epoch": 2.086473131562693,
2373
+ "grad_norm": 0.783029317855835,
2374
+ "learning_rate": 7.004705914739986e-06,
2375
+ "loss": 0.0763,
2376
+ "step": 1690
2377
+ },
2378
+ {
2379
+ "epoch": 2.092649783817171,
2380
+ "grad_norm": 0.714343786239624,
2381
+ "learning_rate": 6.9185146734060465e-06,
2382
+ "loss": 0.0796,
2383
+ "step": 1695
2384
+ },
2385
+ {
2386
+ "epoch": 2.0988264360716493,
2387
+ "grad_norm": 0.9279899597167969,
2388
+ "learning_rate": 6.832697764677881e-06,
2389
+ "loss": 0.0881,
2390
+ "step": 1700
2391
+ },
2392
+ {
2393
+ "epoch": 2.1050030883261273,
2394
+ "grad_norm": 0.7896780371665955,
2395
+ "learning_rate": 6.747259163575648e-06,
2396
+ "loss": 0.0785,
2397
+ "step": 1705
2398
+ },
2399
+ {
2400
+ "epoch": 2.111179740580605,
2401
+ "grad_norm": 1.0813144445419312,
2402
+ "learning_rate": 6.662202827596386e-06,
2403
+ "loss": 0.0791,
2404
+ "step": 1710
2405
+ },
2406
+ {
2407
+ "epoch": 2.1173563928350836,
2408
+ "grad_norm": 0.8622260689735413,
2409
+ "learning_rate": 6.577532696530678e-06,
2410
+ "loss": 0.0855,
2411
+ "step": 1715
2412
+ },
2413
+ {
2414
+ "epoch": 2.1235330450895615,
2415
+ "grad_norm": 0.8274332284927368,
2416
+ "learning_rate": 6.4932526922801935e-06,
2417
+ "loss": 0.0928,
2418
+ "step": 1720
2419
+ },
2420
+ {
2421
+ "epoch": 2.1297096973440395,
2422
+ "grad_norm": 0.8374325633049011,
2423
+ "learning_rate": 6.409366718676031e-06,
2424
+ "loss": 0.0834,
2425
+ "step": 1725
2426
+ },
2427
+ {
2428
+ "epoch": 2.1358863495985174,
2429
+ "grad_norm": 0.9345085024833679,
2430
+ "learning_rate": 6.325878661297845e-06,
2431
+ "loss": 0.0782,
2432
+ "step": 1730
2433
+ },
2434
+ {
2435
+ "epoch": 2.142063001852996,
2436
+ "grad_norm": 0.8159971833229065,
2437
+ "learning_rate": 6.242792387293919e-06,
2438
+ "loss": 0.0737,
2439
+ "step": 1735
2440
+ },
2441
+ {
2442
+ "epoch": 2.1482396541074738,
2443
+ "grad_norm": 0.7708956599235535,
2444
+ "learning_rate": 6.1601117452020125e-06,
2445
+ "loss": 0.0746,
2446
+ "step": 1740
2447
+ },
2448
+ {
2449
+ "epoch": 2.1544163063619517,
2450
+ "grad_norm": 0.8981457948684692,
2451
+ "learning_rate": 6.077840564771116e-06,
2452
+ "loss": 0.0851,
2453
+ "step": 1745
2454
+ },
2455
+ {
2456
+ "epoch": 2.16059295861643,
2457
+ "grad_norm": 0.8107654452323914,
2458
+ "learning_rate": 5.995982656784021e-06,
2459
+ "loss": 0.0862,
2460
+ "step": 1750
2461
+ },
2462
+ {
2463
+ "epoch": 2.166769610870908,
2464
+ "grad_norm": 0.7114669680595398,
2465
+ "learning_rate": 5.914541812880865e-06,
2466
+ "loss": 0.0801,
2467
+ "step": 1755
2468
+ },
2469
+ {
2470
+ "epoch": 2.172946263125386,
2471
+ "grad_norm": 0.6949590444564819,
2472
+ "learning_rate": 5.833521805383451e-06,
2473
+ "loss": 0.0799,
2474
+ "step": 1760
2475
+ },
2476
+ {
2477
+ "epoch": 2.179122915379864,
2478
+ "grad_norm": 0.9366462230682373,
2479
+ "learning_rate": 5.752926387120527e-06,
2480
+ "loss": 0.0755,
2481
+ "step": 1765
2482
+ },
2483
+ {
2484
+ "epoch": 2.1852995676343423,
2485
+ "grad_norm": 0.6815227270126343,
2486
+ "learning_rate": 5.672759291253974e-06,
2487
+ "loss": 0.0745,
2488
+ "step": 1770
2489
+ },
2490
+ {
2491
+ "epoch": 2.1914762198888202,
2492
+ "grad_norm": 0.6726056337356567,
2493
+ "learning_rate": 5.593024231105872e-06,
2494
+ "loss": 0.0767,
2495
+ "step": 1775
2496
+ },
2497
+ {
2498
+ "epoch": 2.197652872143298,
2499
+ "grad_norm": 0.749618411064148,
2500
+ "learning_rate": 5.513724899986501e-06,
2501
+ "loss": 0.0647,
2502
+ "step": 1780
2503
+ },
2504
+ {
2505
+ "epoch": 2.2038295243977766,
2506
+ "grad_norm": 0.7344475388526917,
2507
+ "learning_rate": 5.434864971023269e-06,
2508
+ "loss": 0.084,
2509
+ "step": 1785
2510
+ },
2511
+ {
2512
+ "epoch": 2.2100061766522545,
2513
+ "grad_norm": 0.8653593063354492,
2514
+ "learning_rate": 5.356448096990575e-06,
2515
+ "loss": 0.0747,
2516
+ "step": 1790
2517
+ },
2518
+ {
2519
+ "epoch": 2.2161828289067325,
2520
+ "grad_norm": 0.8198117017745972,
2521
+ "learning_rate": 5.278477910140615e-06,
2522
+ "loss": 0.0742,
2523
+ "step": 1795
2524
+ },
2525
+ {
2526
+ "epoch": 2.2223594811612104,
2527
+ "grad_norm": 0.7100383639335632,
2528
+ "learning_rate": 5.200958022035119e-06,
2529
+ "loss": 0.0745,
2530
+ "step": 1800
2531
+ },
2532
+ {
2533
+ "epoch": 2.228536133415689,
2534
+ "grad_norm": 0.601378858089447,
2535
+ "learning_rate": 5.123892023378098e-06,
2536
+ "loss": 0.0737,
2537
+ "step": 1805
2538
+ },
2539
+ {
2540
+ "epoch": 2.2347127856701667,
2541
+ "grad_norm": 0.6688680648803711,
2542
+ "learning_rate": 5.047283483849497e-06,
2543
+ "loss": 0.0693,
2544
+ "step": 1810
2545
+ },
2546
+ {
2547
+ "epoch": 2.2408894379246447,
2548
+ "grad_norm": 0.6725996732711792,
2549
+ "learning_rate": 4.971135951939853e-06,
2550
+ "loss": 0.0732,
2551
+ "step": 1815
2552
+ },
2553
+ {
2554
+ "epoch": 2.247066090179123,
2555
+ "grad_norm": 0.7726414799690247,
2556
+ "learning_rate": 4.895452954785936e-06,
2557
+ "loss": 0.0763,
2558
+ "step": 1820
2559
+ },
2560
+ {
2561
+ "epoch": 2.253242742433601,
2562
+ "grad_norm": 0.7682061195373535,
2563
+ "learning_rate": 4.820237998007369e-06,
2564
+ "loss": 0.0755,
2565
+ "step": 1825
2566
+ },
2567
+ {
2568
+ "epoch": 2.259419394688079,
2569
+ "grad_norm": 0.6193221211433411,
2570
+ "learning_rate": 4.745494565544252e-06,
2571
+ "loss": 0.0812,
2572
+ "step": 1830
2573
+ },
2574
+ {
2575
+ "epoch": 2.265596046942557,
2576
+ "grad_norm": 0.8341145515441895,
2577
+ "learning_rate": 4.671226119495768e-06,
2578
+ "loss": 0.0814,
2579
+ "step": 1835
2580
+ },
2581
+ {
2582
+ "epoch": 2.2717726991970353,
2583
+ "grad_norm": 0.7899558544158936,
2584
+ "learning_rate": 4.597436099959849e-06,
2585
+ "loss": 0.083,
2586
+ "step": 1840
2587
+ },
2588
+ {
2589
+ "epoch": 2.2779493514515132,
2590
+ "grad_norm": 0.7506449818611145,
2591
+ "learning_rate": 4.5241279248738285e-06,
2592
+ "loss": 0.0791,
2593
+ "step": 1845
2594
+ },
2595
+ {
2596
+ "epoch": 2.284126003705991,
2597
+ "grad_norm": 0.7403045892715454,
2598
+ "learning_rate": 4.451304989856083e-06,
2599
+ "loss": 0.0807,
2600
+ "step": 1850
2601
+ },
2602
+ {
2603
+ "epoch": 2.2903026559604696,
2604
+ "grad_norm": 0.7014782428741455,
2605
+ "learning_rate": 4.378970668048795e-06,
2606
+ "loss": 0.0761,
2607
+ "step": 1855
2608
+ },
2609
+ {
2610
+ "epoch": 2.2964793082149475,
2611
+ "grad_norm": 0.7414482235908508,
2612
+ "learning_rate": 4.307128309961694e-06,
2613
+ "loss": 0.0798,
2614
+ "step": 1860
2615
+ },
2616
+ {
2617
+ "epoch": 2.3026559604694254,
2618
+ "grad_norm": 0.7590753436088562,
2619
+ "learning_rate": 4.235781243316845e-06,
2620
+ "loss": 0.0723,
2621
+ "step": 1865
2622
+ },
2623
+ {
2624
+ "epoch": 2.308832612723904,
2625
+ "grad_norm": 0.6324246525764465,
2626
+ "learning_rate": 4.164932772894532e-06,
2627
+ "loss": 0.0709,
2628
+ "step": 1870
2629
+ },
2630
+ {
2631
+ "epoch": 2.3150092649783818,
2632
+ "grad_norm": 0.8996341228485107,
2633
+ "learning_rate": 4.094586180380186e-06,
2634
+ "loss": 0.0744,
2635
+ "step": 1875
2636
+ },
2637
+ {
2638
+ "epoch": 2.3211859172328597,
2639
+ "grad_norm": 0.7234885692596436,
2640
+ "learning_rate": 4.024744724212335e-06,
2641
+ "loss": 0.0735,
2642
+ "step": 1880
2643
+ },
2644
+ {
2645
+ "epoch": 2.327362569487338,
2646
+ "grad_norm": 0.8266810178756714,
2647
+ "learning_rate": 3.955411639431728e-06,
2648
+ "loss": 0.0743,
2649
+ "step": 1885
2650
+ },
2651
+ {
2652
+ "epoch": 2.333539221741816,
2653
+ "grad_norm": 0.6022233366966248,
2654
+ "learning_rate": 3.886590137531454e-06,
2655
+ "loss": 0.0684,
2656
+ "step": 1890
2657
+ },
2658
+ {
2659
+ "epoch": 2.339715873996294,
2660
+ "grad_norm": 0.5976939797401428,
2661
+ "learning_rate": 3.818283406308205e-06,
2662
+ "loss": 0.0705,
2663
+ "step": 1895
2664
+ },
2665
+ {
2666
+ "epoch": 2.345892526250772,
2667
+ "grad_norm": 0.6941046118736267,
2668
+ "learning_rate": 3.7504946097145885e-06,
2669
+ "loss": 0.068,
2670
+ "step": 1900
2671
+ },
2672
+ {
2673
+ "epoch": 2.3520691785052503,
2674
+ "grad_norm": 0.6822929978370667,
2675
+ "learning_rate": 3.6832268877126236e-06,
2676
+ "loss": 0.0714,
2677
+ "step": 1905
2678
+ },
2679
+ {
2680
+ "epoch": 2.3582458307597283,
2681
+ "grad_norm": 0.7743008136749268,
2682
+ "learning_rate": 3.616483356128256e-06,
2683
+ "loss": 0.0774,
2684
+ "step": 1910
2685
+ },
2686
+ {
2687
+ "epoch": 2.364422483014206,
2688
+ "grad_norm": 0.6219515800476074,
2689
+ "learning_rate": 3.550267106507038e-06,
2690
+ "loss": 0.0615,
2691
+ "step": 1915
2692
+ },
2693
+ {
2694
+ "epoch": 2.3705991352686846,
2695
+ "grad_norm": 0.7663429975509644,
2696
+ "learning_rate": 3.484581205970955e-06,
2697
+ "loss": 0.0703,
2698
+ "step": 1920
2699
+ },
2700
+ {
2701
+ "epoch": 2.3767757875231625,
2702
+ "grad_norm": 0.685067355632782,
2703
+ "learning_rate": 3.419428697076333e-06,
2704
+ "loss": 0.0769,
2705
+ "step": 1925
2706
+ },
2707
+ {
2708
+ "epoch": 2.3829524397776405,
2709
+ "grad_norm": 0.6846317052841187,
2710
+ "learning_rate": 3.354812597672914e-06,
2711
+ "loss": 0.069,
2712
+ "step": 1930
2713
+ },
2714
+ {
2715
+ "epoch": 2.3891290920321184,
2716
+ "grad_norm": 0.5359989404678345,
2717
+ "learning_rate": 3.290735900764077e-06,
2718
+ "loss": 0.0719,
2719
+ "step": 1935
2720
+ },
2721
+ {
2722
+ "epoch": 2.395305744286597,
2723
+ "grad_norm": 0.6229488849639893,
2724
+ "learning_rate": 3.227201574368193e-06,
2725
+ "loss": 0.08,
2726
+ "step": 1940
2727
+ },
2728
+ {
2729
+ "epoch": 2.4014823965410748,
2730
+ "grad_norm": 0.60844486951828,
2731
+ "learning_rate": 3.164212561381156e-06,
2732
+ "loss": 0.0714,
2733
+ "step": 1945
2734
+ },
2735
+ {
2736
+ "epoch": 2.4076590487955527,
2737
+ "grad_norm": 0.5528505444526672,
2738
+ "learning_rate": 3.1017717794400547e-06,
2739
+ "loss": 0.0651,
2740
+ "step": 1950
2741
+ },
2742
+ {
2743
+ "epoch": 2.413835701050031,
2744
+ "grad_norm": 0.7302167415618896,
2745
+ "learning_rate": 3.0398821207880433e-06,
2746
+ "loss": 0.0706,
2747
+ "step": 1955
2748
+ },
2749
+ {
2750
+ "epoch": 2.420012353304509,
2751
+ "grad_norm": 0.6206949949264526,
2752
+ "learning_rate": 2.9785464521403693e-06,
2753
+ "loss": 0.0737,
2754
+ "step": 1960
2755
+ },
2756
+ {
2757
+ "epoch": 2.426189005558987,
2758
+ "grad_norm": 0.63639897108078,
2759
+ "learning_rate": 2.9177676145515856e-06,
2760
+ "loss": 0.0701,
2761
+ "step": 1965
2762
+ },
2763
+ {
2764
+ "epoch": 2.432365657813465,
2765
+ "grad_norm": 0.6214491724967957,
2766
+ "learning_rate": 2.8575484232839527e-06,
2767
+ "loss": 0.0723,
2768
+ "step": 1970
2769
+ },
2770
+ {
2771
+ "epoch": 2.4385423100679433,
2772
+ "grad_norm": 0.5465853214263916,
2773
+ "learning_rate": 2.7978916676770363e-06,
2774
+ "loss": 0.0629,
2775
+ "step": 1975
2776
+ },
2777
+ {
2778
+ "epoch": 2.4447189623224213,
2779
+ "grad_norm": 0.6244093775749207,
2780
+ "learning_rate": 2.7388001110185145e-06,
2781
+ "loss": 0.0701,
2782
+ "step": 1980
2783
+ },
2784
+ {
2785
+ "epoch": 2.450895614576899,
2786
+ "grad_norm": 0.6304371953010559,
2787
+ "learning_rate": 2.680276490416162e-06,
2788
+ "loss": 0.0555,
2789
+ "step": 1985
2790
+ },
2791
+ {
2792
+ "epoch": 2.4570722668313776,
2793
+ "grad_norm": 0.543263852596283,
2794
+ "learning_rate": 2.622323516671092e-06,
2795
+ "loss": 0.0622,
2796
+ "step": 1990
2797
+ },
2798
+ {
2799
+ "epoch": 2.4632489190858555,
2800
+ "grad_norm": 0.6078624725341797,
2801
+ "learning_rate": 2.5649438741521812e-06,
2802
+ "loss": 0.0786,
2803
+ "step": 1995
2804
+ },
2805
+ {
2806
+ "epoch": 2.4694255713403335,
2807
+ "grad_norm": 0.6344765424728394,
2808
+ "learning_rate": 2.5081402206717293e-06,
2809
+ "loss": 0.0737,
2810
+ "step": 2000
2811
+ },
2812
+ {
2813
+ "epoch": 2.4756022235948114,
2814
+ "grad_norm": 0.6288312673568726,
2815
+ "learning_rate": 2.45191518736235e-06,
2816
+ "loss": 0.0678,
2817
+ "step": 2005
2818
+ },
2819
+ {
2820
+ "epoch": 2.48177887584929,
2821
+ "grad_norm": 0.770402193069458,
2822
+ "learning_rate": 2.396271378555109e-06,
2823
+ "loss": 0.0634,
2824
+ "step": 2010
2825
+ },
2826
+ {
2827
+ "epoch": 2.4879555281037677,
2828
+ "grad_norm": 0.5953220129013062,
2829
+ "learning_rate": 2.341211371658861e-06,
2830
+ "loss": 0.0623,
2831
+ "step": 2015
2832
+ },
2833
+ {
2834
+ "epoch": 2.4941321803582457,
2835
+ "grad_norm": 0.7179603576660156,
2836
+ "learning_rate": 2.2867377170409066e-06,
2837
+ "loss": 0.0658,
2838
+ "step": 2020
2839
+ },
2840
+ {
2841
+ "epoch": 2.500308832612724,
2842
+ "grad_norm": 0.6675596833229065,
2843
+ "learning_rate": 2.232852937908831e-06,
2844
+ "loss": 0.0719,
2845
+ "step": 2025
2846
+ },
2847
+ {
2848
+ "epoch": 2.506485484867202,
2849
+ "grad_norm": 0.572619616985321,
2850
+ "learning_rate": 2.179559530193638e-06,
2851
+ "loss": 0.0664,
2852
+ "step": 2030
2853
+ },
2854
+ {
2855
+ "epoch": 2.51266213712168,
2856
+ "grad_norm": 0.6647475957870483,
2857
+ "learning_rate": 2.126859962434137e-06,
2858
+ "loss": 0.0654,
2859
+ "step": 2035
2860
+ },
2861
+ {
2862
+ "epoch": 2.518838789376158,
2863
+ "grad_norm": 0.551464855670929,
2864
+ "learning_rate": 2.0747566756626073e-06,
2865
+ "loss": 0.0635,
2866
+ "step": 2040
2867
+ },
2868
+ {
2869
+ "epoch": 2.5250154416306363,
2870
+ "grad_norm": 0.7556998133659363,
2871
+ "learning_rate": 2.023252083291723e-06,
2872
+ "loss": 0.0725,
2873
+ "step": 2045
2874
+ },
2875
+ {
2876
+ "epoch": 2.5311920938851142,
2877
+ "grad_norm": 0.5222184658050537,
2878
+ "learning_rate": 1.9723485710027595e-06,
2879
+ "loss": 0.0597,
2880
+ "step": 2050
2881
+ },
2882
+ {
2883
+ "epoch": 2.537368746139592,
2884
+ "grad_norm": 0.5315612554550171,
2885
+ "learning_rate": 1.922048496635104e-06,
2886
+ "loss": 0.0708,
2887
+ "step": 2055
2888
+ },
2889
+ {
2890
+ "epoch": 2.5435453983940706,
2891
+ "grad_norm": 0.44711750745773315,
2892
+ "learning_rate": 1.8723541900770414e-06,
2893
+ "loss": 0.0586,
2894
+ "step": 2060
2895
+ },
2896
+ {
2897
+ "epoch": 2.5497220506485485,
2898
+ "grad_norm": 0.567542552947998,
2899
+ "learning_rate": 1.8232679531578044e-06,
2900
+ "loss": 0.0607,
2901
+ "step": 2065
2902
+ },
2903
+ {
2904
+ "epoch": 2.5558987029030265,
2905
+ "grad_norm": 0.6562327742576599,
2906
+ "learning_rate": 1.7747920595409928e-06,
2907
+ "loss": 0.0615,
2908
+ "step": 2070
2909
+ },
2910
+ {
2911
+ "epoch": 2.5620753551575044,
2912
+ "grad_norm": 0.6160925626754761,
2913
+ "learning_rate": 1.7269287546192347e-06,
2914
+ "loss": 0.0602,
2915
+ "step": 2075
2916
+ },
2917
+ {
2918
+ "epoch": 2.568252007411983,
2919
+ "grad_norm": 0.7236652374267578,
2920
+ "learning_rate": 1.6796802554101863e-06,
2921
+ "loss": 0.068,
2922
+ "step": 2080
2923
+ },
2924
+ {
2925
+ "epoch": 2.5744286596664607,
2926
+ "grad_norm": 0.7122112512588501,
2927
+ "learning_rate": 1.633048750453832e-06,
2928
+ "loss": 0.0686,
2929
+ "step": 2085
2930
+ },
2931
+ {
2932
+ "epoch": 2.580605311920939,
2933
+ "grad_norm": 0.5421215295791626,
2934
+ "learning_rate": 1.5870363997111354e-06,
2935
+ "loss": 0.0602,
2936
+ "step": 2090
2937
+ },
2938
+ {
2939
+ "epoch": 2.586781964175417,
2940
+ "grad_norm": 0.5899111032485962,
2941
+ "learning_rate": 1.5416453344639654e-06,
2942
+ "loss": 0.0658,
2943
+ "step": 2095
2944
+ },
2945
+ {
2946
+ "epoch": 2.592958616429895,
2947
+ "grad_norm": 0.6031246781349182,
2948
+ "learning_rate": 1.4968776572163794e-06,
2949
+ "loss": 0.0658,
2950
+ "step": 2100
2951
+ },
2952
+ {
2953
+ "epoch": 2.599135268684373,
2954
+ "grad_norm": 0.5804592370986938,
2955
+ "learning_rate": 1.4527354415972465e-06,
2956
+ "loss": 0.0672,
2957
+ "step": 2105
2958
+ },
2959
+ {
2960
+ "epoch": 2.605311920938851,
2961
+ "grad_norm": 0.6297616362571716,
2962
+ "learning_rate": 1.4092207322641964e-06,
2963
+ "loss": 0.0581,
2964
+ "step": 2110
2965
+ },
2966
+ {
2967
+ "epoch": 2.6114885731933293,
2968
+ "grad_norm": 0.5267701148986816,
2969
+ "learning_rate": 1.366335544808901e-06,
2970
+ "loss": 0.0639,
2971
+ "step": 2115
2972
+ },
2973
+ {
2974
+ "epoch": 2.617665225447807,
2975
+ "grad_norm": 0.6203505992889404,
2976
+ "learning_rate": 1.3240818656637188e-06,
2977
+ "loss": 0.0689,
2978
+ "step": 2120
2979
+ },
2980
+ {
2981
+ "epoch": 2.6238418777022856,
2982
+ "grad_norm": 0.5993887186050415,
2983
+ "learning_rate": 1.2824616520096871e-06,
2984
+ "loss": 0.0658,
2985
+ "step": 2125
2986
+ },
2987
+ {
2988
+ "epoch": 2.6300185299567636,
2989
+ "grad_norm": 0.5098360180854797,
2990
+ "learning_rate": 1.241476831685861e-06,
2991
+ "loss": 0.0651,
2992
+ "step": 2130
2993
+ },
2994
+ {
2995
+ "epoch": 2.6361951822112415,
2996
+ "grad_norm": 1.063259482383728,
2997
+ "learning_rate": 1.2011293031000131e-06,
2998
+ "loss": 0.0601,
2999
+ "step": 2135
3000
+ },
3001
+ {
3002
+ "epoch": 2.6423718344657194,
3003
+ "grad_norm": 0.6282554268836975,
3004
+ "learning_rate": 1.1614209351407106e-06,
3005
+ "loss": 0.0634,
3006
+ "step": 2140
3007
+ },
3008
+ {
3009
+ "epoch": 2.6485484867201974,
3010
+ "grad_norm": 0.6264044046401978,
3011
+ "learning_rate": 1.1223535670907397e-06,
3012
+ "loss": 0.0613,
3013
+ "step": 2145
3014
+ },
3015
+ {
3016
+ "epoch": 2.6547251389746758,
3017
+ "grad_norm": 0.7208899855613708,
3018
+ "learning_rate": 1.0839290085419134e-06,
3019
+ "loss": 0.0673,
3020
+ "step": 2150
3021
+ },
3022
+ {
3023
+ "epoch": 2.6609017912291537,
3024
+ "grad_norm": 0.6089773774147034,
3025
+ "learning_rate": 1.0461490393112505e-06,
3026
+ "loss": 0.0692,
3027
+ "step": 2155
3028
+ },
3029
+ {
3030
+ "epoch": 2.667078443483632,
3031
+ "grad_norm": 0.5348822474479675,
3032
+ "learning_rate": 1.0090154093585396e-06,
3033
+ "loss": 0.0574,
3034
+ "step": 2160
3035
+ },
3036
+ {
3037
+ "epoch": 2.67325509573811,
3038
+ "grad_norm": 0.5450313091278076,
3039
+ "learning_rate": 9.725298387052695e-07,
3040
+ "loss": 0.068,
3041
+ "step": 2165
3042
+ },
3043
+ {
3044
+ "epoch": 2.679431747992588,
3045
+ "grad_norm": 0.6886411309242249,
3046
+ "learning_rate": 9.366940173549721e-07,
3047
+ "loss": 0.0629,
3048
+ "step": 2170
3049
+ },
3050
+ {
3051
+ "epoch": 2.685608400247066,
3052
+ "grad_norm": 0.6111475229263306,
3053
+ "learning_rate": 9.015096052149374e-07,
3054
+ "loss": 0.0631,
3055
+ "step": 2175
3056
+ },
3057
+ {
3058
+ "epoch": 2.6917850525015443,
3059
+ "grad_norm": 0.47397884726524353,
3060
+ "learning_rate": 8.669782320193198e-07,
3061
+ "loss": 0.0619,
3062
+ "step": 2180
3063
+ },
3064
+ {
3065
+ "epoch": 2.6979617047560223,
3066
+ "grad_norm": 0.6356688141822815,
3067
+ "learning_rate": 8.331014972536599e-07,
3068
+ "loss": 0.0655,
3069
+ "step": 2185
3070
+ },
3071
+ {
3072
+ "epoch": 2.7041383570105,
3073
+ "grad_norm": 0.7169201970100403,
3074
+ "learning_rate": 7.998809700807891e-07,
3075
+ "loss": 0.0569,
3076
+ "step": 2190
3077
+ },
3078
+ {
3079
+ "epoch": 2.7103150092649786,
3080
+ "grad_norm": 0.7548785209655762,
3081
+ "learning_rate": 7.673181892681508e-07,
3082
+ "loss": 0.0732,
3083
+ "step": 2195
3084
+ },
3085
+ {
3086
+ "epoch": 2.7164916615194565,
3087
+ "grad_norm": 0.5386422872543335,
3088
+ "learning_rate": 7.35414663116516e-07,
3089
+ "loss": 0.063,
3090
+ "step": 2200
3091
+ },
3092
+ {
3093
+ "epoch": 2.7226683137739345,
3094
+ "grad_norm": 0.821855366230011,
3095
+ "learning_rate": 7.041718693901355e-07,
3096
+ "loss": 0.0566,
3097
+ "step": 2205
3098
+ },
3099
+ {
3100
+ "epoch": 2.7288449660284124,
3101
+ "grad_norm": 0.6750683188438416,
3102
+ "learning_rate": 6.735912552482787e-07,
3103
+ "loss": 0.0647,
3104
+ "step": 2210
3105
+ },
3106
+ {
3107
+ "epoch": 2.735021618282891,
3108
+ "grad_norm": 0.606149435043335,
3109
+ "learning_rate": 6.436742371781995e-07,
3110
+ "loss": 0.0668,
3111
+ "step": 2215
3112
+ },
3113
+ {
3114
+ "epoch": 2.7411982705373688,
3115
+ "grad_norm": 0.5885270833969116,
3116
+ "learning_rate": 6.144222009295314e-07,
3117
+ "loss": 0.0593,
3118
+ "step": 2220
3119
+ },
3120
+ {
3121
+ "epoch": 2.7473749227918467,
3122
+ "grad_norm": 0.799705982208252,
3123
+ "learning_rate": 5.858365014500983e-07,
3124
+ "loss": 0.0637,
3125
+ "step": 2225
3126
+ },
3127
+ {
3128
+ "epoch": 2.753551575046325,
3129
+ "grad_norm": 0.6732608079910278,
3130
+ "learning_rate": 5.579184628231548e-07,
3131
+ "loss": 0.0776,
3132
+ "step": 2230
3133
+ },
3134
+ {
3135
+ "epoch": 2.759728227300803,
3136
+ "grad_norm": 0.5880258083343506,
3137
+ "learning_rate": 5.306693782060457e-07,
3138
+ "loss": 0.0667,
3139
+ "step": 2235
3140
+ },
3141
+ {
3142
+ "epoch": 2.765904879555281,
3143
+ "grad_norm": 0.5307620167732239,
3144
+ "learning_rate": 5.040905097703246e-07,
3145
+ "loss": 0.0613,
3146
+ "step": 2240
3147
+ },
3148
+ {
3149
+ "epoch": 2.772081531809759,
3150
+ "grad_norm": 0.5822154879570007,
3151
+ "learning_rate": 4.781830886432737e-07,
3152
+ "loss": 0.0601,
3153
+ "step": 2245
3154
+ },
3155
+ {
3156
+ "epoch": 2.7782581840642373,
3157
+ "grad_norm": 0.6783782839775085,
3158
+ "learning_rate": 4.5294831485088453e-07,
3159
+ "loss": 0.0716,
3160
+ "step": 2250
3161
+ },
3162
+ {
3163
+ "epoch": 2.7844348363187152,
3164
+ "grad_norm": 0.45106980204582214,
3165
+ "learning_rate": 4.2838735726227876e-07,
3166
+ "loss": 0.0677,
3167
+ "step": 2255
3168
+ },
3169
+ {
3170
+ "epoch": 2.790611488573193,
3171
+ "grad_norm": 0.6480432152748108,
3172
+ "learning_rate": 4.0450135353555573e-07,
3173
+ "loss": 0.0689,
3174
+ "step": 2260
3175
+ },
3176
+ {
3177
+ "epoch": 2.7967881408276716,
3178
+ "grad_norm": 0.6420740485191345,
3179
+ "learning_rate": 3.8129141006510494e-07,
3180
+ "loss": 0.0602,
3181
+ "step": 2265
3182
+ },
3183
+ {
3184
+ "epoch": 2.8029647930821495,
3185
+ "grad_norm": 0.5628190040588379,
3186
+ "learning_rate": 3.587586019303535e-07,
3187
+ "loss": 0.0592,
3188
+ "step": 2270
3189
+ },
3190
+ {
3191
+ "epoch": 2.8091414453366275,
3192
+ "grad_norm": 0.6839898824691772,
3193
+ "learning_rate": 3.369039728459711e-07,
3194
+ "loss": 0.0629,
3195
+ "step": 2275
3196
+ },
3197
+ {
3198
+ "epoch": 2.8153180975911054,
3199
+ "grad_norm": 0.7047255039215088,
3200
+ "learning_rate": 3.1572853511352694e-07,
3201
+ "loss": 0.0608,
3202
+ "step": 2280
3203
+ },
3204
+ {
3205
+ "epoch": 2.821494749845584,
3206
+ "grad_norm": 0.519675076007843,
3207
+ "learning_rate": 2.9523326957458907e-07,
3208
+ "loss": 0.0665,
3209
+ "step": 2285
3210
+ },
3211
+ {
3212
+ "epoch": 2.8276714021000617,
3213
+ "grad_norm": 0.5927917957305908,
3214
+ "learning_rate": 2.754191255653088e-07,
3215
+ "loss": 0.0667,
3216
+ "step": 2290
3217
+ },
3218
+ {
3219
+ "epoch": 2.8338480543545397,
3220
+ "grad_norm": 0.5696297287940979,
3221
+ "learning_rate": 2.5628702087243606e-07,
3222
+ "loss": 0.063,
3223
+ "step": 2295
3224
+ },
3225
+ {
3226
+ "epoch": 2.840024706609018,
3227
+ "grad_norm": 0.500535786151886,
3228
+ "learning_rate": 2.3783784169081336e-07,
3229
+ "loss": 0.0638,
3230
+ "step": 2300
3231
+ },
3232
+ {
3233
+ "epoch": 2.846201358863496,
3234
+ "grad_norm": 0.5720264911651611,
3235
+ "learning_rate": 2.2007244258232194e-07,
3236
+ "loss": 0.0569,
3237
+ "step": 2305
3238
+ },
3239
+ {
3240
+ "epoch": 2.852378011117974,
3241
+ "grad_norm": 0.5841343998908997,
3242
+ "learning_rate": 2.029916464363052e-07,
3243
+ "loss": 0.0608,
3244
+ "step": 2310
3245
+ },
3246
+ {
3247
+ "epoch": 2.858554663372452,
3248
+ "grad_norm": 0.6318969130516052,
3249
+ "learning_rate": 1.865962444314473e-07,
3250
+ "loss": 0.0686,
3251
+ "step": 2315
3252
+ },
3253
+ {
3254
+ "epoch": 2.8647313156269303,
3255
+ "grad_norm": 0.7228346467018127,
3256
+ "learning_rate": 1.7088699599912605e-07,
3257
+ "loss": 0.0701,
3258
+ "step": 2320
3259
+ },
3260
+ {
3261
+ "epoch": 2.8709079678814082,
3262
+ "grad_norm": 0.6243155598640442,
3263
+ "learning_rate": 1.5586462878824426e-07,
3264
+ "loss": 0.0715,
3265
+ "step": 2325
3266
+ },
3267
+ {
3268
+ "epoch": 2.8770846201358866,
3269
+ "grad_norm": 0.5605066418647766,
3270
+ "learning_rate": 1.415298386315117e-07,
3271
+ "loss": 0.0577,
3272
+ "step": 2330
3273
+ },
3274
+ {
3275
+ "epoch": 2.8832612723903646,
3276
+ "grad_norm": 0.6332428455352783,
3277
+ "learning_rate": 1.2788328951322758e-07,
3278
+ "loss": 0.0728,
3279
+ "step": 2335
3280
+ },
3281
+ {
3282
+ "epoch": 2.8894379246448425,
3283
+ "grad_norm": 0.6142319440841675,
3284
+ "learning_rate": 1.1492561353851683e-07,
3285
+ "loss": 0.0713,
3286
+ "step": 2340
3287
+ },
3288
+ {
3289
+ "epoch": 2.8956145768993204,
3290
+ "grad_norm": 0.6322025656700134,
3291
+ "learning_rate": 1.0265741090405179e-07,
3292
+ "loss": 0.0673,
3293
+ "step": 2345
3294
+ },
3295
+ {
3296
+ "epoch": 2.9017912291537984,
3297
+ "grad_norm": 0.5626116991043091,
3298
+ "learning_rate": 9.107924987025618e-08,
3299
+ "loss": 0.0649,
3300
+ "step": 2350
3301
+ },
3302
+ {
3303
+ "epoch": 2.9079678814082768,
3304
+ "grad_norm": 0.5709447264671326,
3305
+ "learning_rate": 8.019166673497613e-08,
3306
+ "loss": 0.0608,
3307
+ "step": 2355
3308
+ },
3309
+ {
3310
+ "epoch": 2.9141445336627547,
3311
+ "grad_norm": 0.5484122037887573,
3312
+ "learning_rate": 6.999516580864507e-08,
3313
+ "loss": 0.0612,
3314
+ "step": 2360
3315
+ },
3316
+ {
3317
+ "epoch": 2.920321185917233,
3318
+ "grad_norm": 0.61465984582901,
3319
+ "learning_rate": 6.049021939092236e-08,
3320
+ "loss": 0.0547,
3321
+ "step": 2365
3322
+ },
3323
+ {
3324
+ "epoch": 2.926497838171711,
3325
+ "grad_norm": 0.6099061369895935,
3326
+ "learning_rate": 5.167726774881587e-08,
3327
+ "loss": 0.0649,
3328
+ "step": 2370
3329
+ },
3330
+ {
3331
+ "epoch": 2.932674490426189,
3332
+ "grad_norm": 0.5663135051727295,
3333
+ "learning_rate": 4.355671909628822e-08,
3334
+ "loss": 0.0648,
3335
+ "step": 2375
3336
+ },
3337
+ {
3338
+ "epoch": 2.938851142680667,
3339
+ "grad_norm": 0.6620079874992371,
3340
+ "learning_rate": 3.612894957535029e-08,
3341
+ "loss": 0.067,
3342
+ "step": 2380
3343
+ },
3344
+ {
3345
+ "epoch": 2.945027794935145,
3346
+ "grad_norm": 0.5970832705497742,
3347
+ "learning_rate": 2.9394303238635124e-08,
3348
+ "loss": 0.061,
3349
+ "step": 2385
3350
+ },
3351
+ {
3352
+ "epoch": 2.9512044471896233,
3353
+ "grad_norm": 0.49848443269729614,
3354
+ "learning_rate": 2.335309203346736e-08,
3355
+ "loss": 0.0599,
3356
+ "step": 2390
3357
+ },
3358
+ {
3359
+ "epoch": 2.957381099444101,
3360
+ "grad_norm": 0.5331148505210876,
3361
+ "learning_rate": 1.8005595787408123e-08,
3362
+ "loss": 0.0576,
3363
+ "step": 2395
3364
+ },
3365
+ {
3366
+ "epoch": 2.9635577516985796,
3367
+ "grad_norm": 0.5320906043052673,
3368
+ "learning_rate": 1.335206219529539e-08,
3369
+ "loss": 0.0621,
3370
+ "step": 2400
3371
+ },
3372
+ {
3373
+ "epoch": 2.9697344039530575,
3374
+ "grad_norm": 0.46187058091163635,
3375
+ "learning_rate": 9.39270680777149e-09,
3376
+ "loss": 0.0611,
3377
+ "step": 2405
3378
+ },
3379
+ {
3380
+ "epoch": 2.9759110562075355,
3381
+ "grad_norm": 0.628253161907196,
3382
+ "learning_rate": 6.127713021301107e-09,
3383
+ "loss": 0.0622,
3384
+ "step": 2410
3385
+ },
3386
+ {
3387
+ "epoch": 2.9820877084620134,
3388
+ "grad_norm": 0.6351046562194824,
3389
+ "learning_rate": 3.5572320696697358e-09,
3390
+ "loss": 0.0594,
3391
+ "step": 2415
3392
+ },
3393
+ {
3394
+ "epoch": 2.988264360716492,
3395
+ "grad_norm": 0.6182594299316406,
3396
+ "learning_rate": 1.6813830169826183e-09,
3397
+ "loss": 0.0627,
3398
+ "step": 2420
3399
+ },
3400
+ {
3401
+ "epoch": 2.9944410129709698,
3402
+ "grad_norm": 0.6555151343345642,
3403
+ "learning_rate": 5.002527521541511e-10,
3404
+ "loss": 0.0702,
3405
+ "step": 2425
3406
+ },
3407
+ {
3408
+ "epoch": 3.0,
3409
+ "grad_norm": 0.5810496211051941,
3410
+ "learning_rate": 1.38959848761111e-11,
3411
+ "loss": 0.0565,
3412
+ "step": 2430
3413
+ },
3414
+ {
3415
+ "epoch": 3.0,
3416
+ "step": 2430,
3417
+ "total_flos": 3.277186621681697e+18,
3418
+ "train_loss": 0.3510698379803096,
3419
+ "train_runtime": 1515.331,
3420
+ "train_samples_per_second": 51.27,
3421
+ "train_steps_per_second": 1.604
3422
+ }
3423
+ ],
3424
+ "logging_steps": 5,
3425
+ "max_steps": 2430,
3426
+ "num_input_tokens_seen": 0,
3427
+ "num_train_epochs": 3,
3428
+ "save_steps": 20000,
3429
+ "stateful_callbacks": {
3430
+ "TrainerControl": {
3431
+ "args": {
3432
+ "should_epoch_stop": false,
3433
+ "should_evaluate": false,
3434
+ "should_log": false,
3435
+ "should_save": false,
3436
+ "should_training_stop": false
3437
+ },
3438
+ "attributes": {}
3439
+ }
3440
+ },
3441
+ "total_flos": 3.277186621681697e+18,
3442
+ "train_batch_size": 2,
3443
+ "trial_name": null,
3444
+ "trial_params": null
3445
+ }
instruct/116_128_e3_3e-5/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23c32de21c77185001d82378d203d2258a7f290e8a3352b369c61c7f776aaf6d
3
+ size 8273
instruct/116_128_e3_3e-5/vocab.json ADDED
The diff for this file is too large to render. See raw diff