Hugging Face's logo

RayDu0010
/
tulu10k_klora

Safetensors
Model card Files
xet
Community
RayDu0010 commited on
Commit
2d8040a
·
verified ·
1 Parent(s): 8cd12fa

Upload folder using huggingface_hub

Browse files
Files changed (14) hide show
  1. instruct/134_128_e3_3e-5/adapter_config.json +39 -0
  2. instruct/134_128_e3_3e-5/adapter_model.safetensors +3 -0
  3. instruct/134_128_e3_3e-5/added_tokens.json +9 -0
  4. instruct/134_128_e3_3e-5/all_results.json +9 -0
  5. instruct/134_128_e3_3e-5/chat_template.jinja +62 -0
  6. instruct/134_128_e3_3e-5/config.json +32 -0
  7. instruct/134_128_e3_3e-5/merges.txt +0 -0
  8. instruct/134_128_e3_3e-5/special_tokens_map.json +33 -0
  9. instruct/134_128_e3_3e-5/tokenizer.json +0 -0
  10. instruct/134_128_e3_3e-5/tokenizer_config.json +234 -0
  11. instruct/134_128_e3_3e-5/train_results.json +9 -0
  12. instruct/134_128_e3_3e-5/trainer_state.json +2906 -0
  13. instruct/134_128_e3_3e-5/training_args.bin +3 -0
  14. instruct/134_128_e3_3e-5/vocab.json +0 -0
instruct/134_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
+ "o_proj",
28
+ "v_proj",
29
+ "gate_proj",
30
+ "down_proj",
31
+ "q_proj",
32
+ "k_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/134_128_e3_3e-5/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01eca857bfe8463faf7aedcc768a6258a2d88601e655dbd764fd5d967878c7b7
3
+ size 791751704
instruct/134_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/134_128_e3_3e-5/all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 2.6763548800791347e+18,
4
+ "train_loss": 0.38947603013411913,
5
+ "train_runtime": 1269.3059,
6
+ "train_samples": 21825,
7
+ "train_samples_per_second": 51.583,
8
+ "train_steps_per_second": 1.614
9
+ }
instruct/134_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/134_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/134_128_e3_3e-5/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
instruct/134_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/134_128_e3_3e-5/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
instruct/134_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/134_128_e3_3e-5/train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "total_flos": 2.6763548800791347e+18,
4
+ "train_loss": 0.38947603013411913,
5
+ "train_runtime": 1269.3059,
6
+ "train_samples": 21825,
7
+ "train_samples_per_second": 51.583,
8
+ "train_steps_per_second": 1.614
9
+ }
instruct/134_128_e3_3e-5/trainer_state.json ADDED
@@ -0,0 +1,2906 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 2049,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.007326007326007326,
14
+ "grad_norm": 2.5924038887023926,
15
+ "learning_rate": 1.1650485436893204e-06,
16
+ "loss": 1.6686,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.014652014652014652,
21
+ "grad_norm": 1.9511922597885132,
22
+ "learning_rate": 2.621359223300971e-06,
23
+ "loss": 1.6131,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.02197802197802198,
28
+ "grad_norm": 1.4122815132141113,
29
+ "learning_rate": 4.0776699029126215e-06,
30
+ "loss": 1.5409,
31
+ "step": 15
32
+ },
33
+ {
34
+ "epoch": 0.029304029304029304,
35
+ "grad_norm": 0.8087025284767151,
36
+ "learning_rate": 5.533980582524272e-06,
37
+ "loss": 1.5666,
38
+ "step": 20
39
+ },
40
+ {
41
+ "epoch": 0.03663003663003663,
42
+ "grad_norm": 0.528700590133667,
43
+ "learning_rate": 6.990291262135923e-06,
44
+ "loss": 1.5256,
45
+ "step": 25
46
+ },
47
+ {
48
+ "epoch": 0.04395604395604396,
49
+ "grad_norm": 0.5109856128692627,
50
+ "learning_rate": 8.446601941747573e-06,
51
+ "loss": 1.446,
52
+ "step": 30
53
+ },
54
+ {
55
+ "epoch": 0.05128205128205128,
56
+ "grad_norm": 0.4208224415779114,
57
+ "learning_rate": 9.902912621359224e-06,
58
+ "loss": 1.4432,
59
+ "step": 35
60
+ },
61
+ {
62
+ "epoch": 0.05860805860805861,
63
+ "grad_norm": 0.3898172676563263,
64
+ "learning_rate": 1.1359223300970873e-05,
65
+ "loss": 1.454,
66
+ "step": 40
67
+ },
68
+ {
69
+ "epoch": 0.06593406593406594,
70
+ "grad_norm": 0.3941916823387146,
71
+ "learning_rate": 1.2815533980582524e-05,
72
+ "loss": 1.4551,
73
+ "step": 45
74
+ },
75
+ {
76
+ "epoch": 0.07326007326007326,
77
+ "grad_norm": 0.38917428255081177,
78
+ "learning_rate": 1.4271844660194176e-05,
79
+ "loss": 1.4657,
80
+ "step": 50
81
+ },
82
+ {
83
+ "epoch": 0.08058608058608059,
84
+ "grad_norm": 0.4119425415992737,
85
+ "learning_rate": 1.5728155339805827e-05,
86
+ "loss": 1.4583,
87
+ "step": 55
88
+ },
89
+ {
90
+ "epoch": 0.08791208791208792,
91
+ "grad_norm": 0.5691628456115723,
92
+ "learning_rate": 1.7184466019417476e-05,
93
+ "loss": 1.3904,
94
+ "step": 60
95
+ },
96
+ {
97
+ "epoch": 0.09523809523809523,
98
+ "grad_norm": 0.3929953873157501,
99
+ "learning_rate": 1.8640776699029126e-05,
100
+ "loss": 1.4176,
101
+ "step": 65
102
+ },
103
+ {
104
+ "epoch": 0.10256410256410256,
105
+ "grad_norm": 0.4677892029285431,
106
+ "learning_rate": 2.009708737864078e-05,
107
+ "loss": 1.4467,
108
+ "step": 70
109
+ },
110
+ {
111
+ "epoch": 0.10989010989010989,
112
+ "grad_norm": 0.39517340064048767,
113
+ "learning_rate": 2.1553398058252428e-05,
114
+ "loss": 1.407,
115
+ "step": 75
116
+ },
117
+ {
118
+ "epoch": 0.11721611721611722,
119
+ "grad_norm": 0.45939716696739197,
120
+ "learning_rate": 2.3009708737864078e-05,
121
+ "loss": 1.3859,
122
+ "step": 80
123
+ },
124
+ {
125
+ "epoch": 0.12454212454212454,
126
+ "grad_norm": 0.4543648958206177,
127
+ "learning_rate": 2.4466019417475727e-05,
128
+ "loss": 1.3652,
129
+ "step": 85
130
+ },
131
+ {
132
+ "epoch": 0.13186813186813187,
133
+ "grad_norm": 0.44294703006744385,
134
+ "learning_rate": 2.5922330097087377e-05,
135
+ "loss": 1.4238,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.1391941391941392,
140
+ "grad_norm": 0.5850459933280945,
141
+ "learning_rate": 2.737864077669903e-05,
142
+ "loss": 1.3706,
143
+ "step": 95
144
+ },
145
+ {
146
+ "epoch": 0.14652014652014653,
147
+ "grad_norm": 0.46346747875213623,
148
+ "learning_rate": 2.8834951456310683e-05,
149
+ "loss": 1.3216,
150
+ "step": 100
151
+ },
152
+ {
153
+ "epoch": 0.15384615384615385,
154
+ "grad_norm": 0.5360530018806458,
155
+ "learning_rate": 2.9999980453219208e-05,
156
+ "loss": 1.2954,
157
+ "step": 105
158
+ },
159
+ {
160
+ "epoch": 0.16117216117216118,
161
+ "grad_norm": 0.5150305032730103,
162
+ "learning_rate": 2.999929632124053e-05,
163
+ "loss": 1.2923,
164
+ "step": 110
165
+ },
166
+ {
167
+ "epoch": 0.1684981684981685,
168
+ "grad_norm": 0.518902063369751,
169
+ "learning_rate": 2.9997634901165178e-05,
170
+ "loss": 1.2937,
171
+ "step": 115
172
+ },
173
+ {
174
+ "epoch": 0.17582417582417584,
175
+ "grad_norm": 0.5556808114051819,
176
+ "learning_rate": 2.999499630124398e-05,
177
+ "loss": 1.2989,
178
+ "step": 120
179
+ },
180
+ {
181
+ "epoch": 0.18315018315018314,
182
+ "grad_norm": 0.5337339639663696,
183
+ "learning_rate": 2.9991380693396476e-05,
184
+ "loss": 1.2512,
185
+ "step": 125
186
+ },
187
+ {
188
+ "epoch": 0.19047619047619047,
189
+ "grad_norm": 0.5952735543251038,
190
+ "learning_rate": 2.9986788313199752e-05,
191
+ "loss": 1.2981,
192
+ "step": 130
193
+ },
194
+ {
195
+ "epoch": 0.1978021978021978,
196
+ "grad_norm": 0.5809885859489441,
197
+ "learning_rate": 2.9981219459873085e-05,
198
+ "loss": 1.1767,
199
+ "step": 135
200
+ },
201
+ {
202
+ "epoch": 0.20512820512820512,
203
+ "grad_norm": 0.5680771470069885,
204
+ "learning_rate": 2.9974674496258428e-05,
205
+ "loss": 1.2253,
206
+ "step": 140
207
+ },
208
+ {
209
+ "epoch": 0.21245421245421245,
210
+ "grad_norm": 0.6411072611808777,
211
+ "learning_rate": 2.9967153848796793e-05,
212
+ "loss": 1.2388,
213
+ "step": 145
214
+ },
215
+ {
216
+ "epoch": 0.21978021978021978,
217
+ "grad_norm": 0.6067551970481873,
218
+ "learning_rate": 2.995865800750045e-05,
219
+ "loss": 1.2052,
220
+ "step": 150
221
+ },
222
+ {
223
+ "epoch": 0.2271062271062271,
224
+ "grad_norm": 0.6428530216217041,
225
+ "learning_rate": 2.9949187525921004e-05,
226
+ "loss": 1.1958,
227
+ "step": 155
228
+ },
229
+ {
230
+ "epoch": 0.23443223443223443,
231
+ "grad_norm": 0.6577104926109314,
232
+ "learning_rate": 2.9938743021113326e-05,
233
+ "loss": 1.1862,
234
+ "step": 160
235
+ },
236
+ {
237
+ "epoch": 0.24175824175824176,
238
+ "grad_norm": 0.6279430389404297,
239
+ "learning_rate": 2.9927325173595366e-05,
240
+ "loss": 1.1638,
241
+ "step": 165
242
+ },
243
+ {
244
+ "epoch": 0.2490842490842491,
245
+ "grad_norm": 0.6782162189483643,
246
+ "learning_rate": 2.9914934727303776e-05,
247
+ "loss": 1.2178,
248
+ "step": 170
249
+ },
250
+ {
251
+ "epoch": 0.2564102564102564,
252
+ "grad_norm": 0.7112299799919128,
253
+ "learning_rate": 2.9901572489545482e-05,
254
+ "loss": 1.1074,
255
+ "step": 175
256
+ },
257
+ {
258
+ "epoch": 0.26373626373626374,
259
+ "grad_norm": 0.7001322507858276,
260
+ "learning_rate": 2.9887239330945047e-05,
261
+ "loss": 1.1224,
262
+ "step": 180
263
+ },
264
+ {
265
+ "epoch": 0.27106227106227104,
266
+ "grad_norm": 0.743632435798645,
267
+ "learning_rate": 2.987193618538797e-05,
268
+ "loss": 1.0887,
269
+ "step": 185
270
+ },
271
+ {
272
+ "epoch": 0.2783882783882784,
273
+ "grad_norm": 0.7162584066390991,
274
+ "learning_rate": 2.985566404995983e-05,
275
+ "loss": 1.092,
276
+ "step": 190
277
+ },
278
+ {
279
+ "epoch": 0.2857142857142857,
280
+ "grad_norm": 0.7778493165969849,
281
+ "learning_rate": 2.983842398488131e-05,
282
+ "loss": 1.0873,
283
+ "step": 195
284
+ },
285
+ {
286
+ "epoch": 0.29304029304029305,
287
+ "grad_norm": 0.7742840647697449,
288
+ "learning_rate": 2.982021711343913e-05,
289
+ "loss": 1.052,
290
+ "step": 200
291
+ },
292
+ {
293
+ "epoch": 0.30036630036630035,
294
+ "grad_norm": 0.7497329711914062,
295
+ "learning_rate": 2.9801044621912862e-05,
296
+ "loss": 1.0782,
297
+ "step": 205
298
+ },
299
+ {
300
+ "epoch": 0.3076923076923077,
301
+ "grad_norm": 0.7409875392913818,
302
+ "learning_rate": 2.9780907759497622e-05,
303
+ "loss": 1.0779,
304
+ "step": 210
305
+ },
306
+ {
307
+ "epoch": 0.315018315018315,
308
+ "grad_norm": 0.8911153674125671,
309
+ "learning_rate": 2.9759807838222688e-05,
310
+ "loss": 0.9939,
311
+ "step": 215
312
+ },
313
+ {
314
+ "epoch": 0.32234432234432236,
315
+ "grad_norm": 0.9070504307746887,
316
+ "learning_rate": 2.9737746232866017e-05,
317
+ "loss": 1.0418,
318
+ "step": 220
319
+ },
320
+ {
321
+ "epoch": 0.32967032967032966,
322
+ "grad_norm": 0.8401923775672913,
323
+ "learning_rate": 2.9714724380864656e-05,
324
+ "loss": 1.001,
325
+ "step": 225
326
+ },
327
+ {
328
+ "epoch": 0.336996336996337,
329
+ "grad_norm": 0.7889268398284912,
330
+ "learning_rate": 2.9690743782221108e-05,
331
+ "loss": 0.9902,
332
+ "step": 230
333
+ },
334
+ {
335
+ "epoch": 0.3443223443223443,
336
+ "grad_norm": 0.8856872320175171,
337
+ "learning_rate": 2.9665805999405576e-05,
338
+ "loss": 1.016,
339
+ "step": 235
340
+ },
341
+ {
342
+ "epoch": 0.3516483516483517,
343
+ "grad_norm": 0.8749246001243591,
344
+ "learning_rate": 2.9639912657254172e-05,
345
+ "loss": 0.9601,
346
+ "step": 240
347
+ },
348
+ {
349
+ "epoch": 0.358974358974359,
350
+ "grad_norm": 0.8690497279167175,
351
+ "learning_rate": 2.9613065442863045e-05,
352
+ "loss": 0.9556,
353
+ "step": 245
354
+ },
355
+ {
356
+ "epoch": 0.3663003663003663,
357
+ "grad_norm": 0.9206748604774475,
358
+ "learning_rate": 2.9585266105478458e-05,
359
+ "loss": 0.9286,
360
+ "step": 250
361
+ },
362
+ {
363
+ "epoch": 0.37362637362637363,
364
+ "grad_norm": 0.8753765225410461,
365
+ "learning_rate": 2.9556516456382817e-05,
366
+ "loss": 0.9363,
367
+ "step": 255
368
+ },
369
+ {
370
+ "epoch": 0.38095238095238093,
371
+ "grad_norm": 0.9768513441085815,
372
+ "learning_rate": 2.952681836877665e-05,
373
+ "loss": 0.8721,
374
+ "step": 260
375
+ },
376
+ {
377
+ "epoch": 0.3882783882783883,
378
+ "grad_norm": 0.9605430960655212,
379
+ "learning_rate": 2.9496173777656575e-05,
380
+ "loss": 0.9242,
381
+ "step": 265
382
+ },
383
+ {
384
+ "epoch": 0.3956043956043956,
385
+ "grad_norm": 0.9489908814430237,
386
+ "learning_rate": 2.9464584679689193e-05,
387
+ "loss": 0.8946,
388
+ "step": 270
389
+ },
390
+ {
391
+ "epoch": 0.40293040293040294,
392
+ "grad_norm": 0.9864665865898132,
393
+ "learning_rate": 2.9432053133081026e-05,
394
+ "loss": 0.887,
395
+ "step": 275
396
+ },
397
+ {
398
+ "epoch": 0.41025641025641024,
399
+ "grad_norm": 0.9729050993919373,
400
+ "learning_rate": 2.939858125744439e-05,
401
+ "loss": 0.8813,
402
+ "step": 280
403
+ },
404
+ {
405
+ "epoch": 0.4175824175824176,
406
+ "grad_norm": 0.8731024861335754,
407
+ "learning_rate": 2.9364171233659303e-05,
408
+ "loss": 0.8827,
409
+ "step": 285
410
+ },
411
+ {
412
+ "epoch": 0.4249084249084249,
413
+ "grad_norm": 1.0156772136688232,
414
+ "learning_rate": 2.9328825303731383e-05,
415
+ "loss": 0.861,
416
+ "step": 290
417
+ },
418
+ {
419
+ "epoch": 0.43223443223443225,
420
+ "grad_norm": 0.9875866770744324,
421
+ "learning_rate": 2.9292545770645776e-05,
422
+ "loss": 0.8621,
423
+ "step": 295
424
+ },
425
+ {
426
+ "epoch": 0.43956043956043955,
427
+ "grad_norm": 0.9691270589828491,
428
+ "learning_rate": 2.92553349982171e-05,
429
+ "loss": 0.8823,
430
+ "step": 300
431
+ },
432
+ {
433
+ "epoch": 0.4468864468864469,
434
+ "grad_norm": 1.0761867761611938,
435
+ "learning_rate": 2.9217195410935414e-05,
436
+ "loss": 0.8371,
437
+ "step": 305
438
+ },
439
+ {
440
+ "epoch": 0.4542124542124542,
441
+ "grad_norm": 1.1158421039581299,
442
+ "learning_rate": 2.9178129493808282e-05,
443
+ "loss": 0.8279,
444
+ "step": 310
445
+ },
446
+ {
447
+ "epoch": 0.46153846153846156,
448
+ "grad_norm": 0.9816854596138,
449
+ "learning_rate": 2.913813979219882e-05,
450
+ "loss": 0.7873,
451
+ "step": 315
452
+ },
453
+ {
454
+ "epoch": 0.46886446886446886,
455
+ "grad_norm": 1.1753109693527222,
456
+ "learning_rate": 2.9097228911659894e-05,
457
+ "loss": 0.7989,
458
+ "step": 320
459
+ },
460
+ {
461
+ "epoch": 0.47619047619047616,
462
+ "grad_norm": 1.2213771343231201,
463
+ "learning_rate": 2.905539951776432e-05,
464
+ "loss": 0.8457,
465
+ "step": 325
466
+ },
467
+ {
468
+ "epoch": 0.4835164835164835,
469
+ "grad_norm": 1.0630041360855103,
470
+ "learning_rate": 2.9012654335931202e-05,
471
+ "loss": 0.8152,
472
+ "step": 330
473
+ },
474
+ {
475
+ "epoch": 0.4908424908424908,
476
+ "grad_norm": 1.1241763830184937,
477
+ "learning_rate": 2.896899615124835e-05,
478
+ "loss": 0.7796,
479
+ "step": 335
480
+ },
481
+ {
482
+ "epoch": 0.4981684981684982,
483
+ "grad_norm": 1.094061255455017,
484
+ "learning_rate": 2.892442780829083e-05,
485
+ "loss": 0.789,
486
+ "step": 340
487
+ },
488
+ {
489
+ "epoch": 0.5054945054945055,
490
+ "grad_norm": 1.1185861825942993,
491
+ "learning_rate": 2.88789522109356e-05,
492
+ "loss": 0.7508,
493
+ "step": 345
494
+ },
495
+ {
496
+ "epoch": 0.5128205128205128,
497
+ "grad_norm": 0.9973278045654297,
498
+ "learning_rate": 2.8832572322172325e-05,
499
+ "loss": 0.783,
500
+ "step": 350
501
+ },
502
+ {
503
+ "epoch": 0.5201465201465202,
504
+ "grad_norm": 1.051804542541504,
505
+ "learning_rate": 2.8785291163910317e-05,
506
+ "loss": 0.7951,
507
+ "step": 355
508
+ },
509
+ {
510
+ "epoch": 0.5274725274725275,
511
+ "grad_norm": 1.0658538341522217,
512
+ "learning_rate": 2.873711181678164e-05,
513
+ "loss": 0.7329,
514
+ "step": 360
515
+ },
516
+ {
517
+ "epoch": 0.5347985347985348,
518
+ "grad_norm": 0.9967415928840637,
519
+ "learning_rate": 2.868803741994039e-05,
520
+ "loss": 0.7305,
521
+ "step": 365
522
+ },
523
+ {
524
+ "epoch": 0.5421245421245421,
525
+ "grad_norm": 1.0558112859725952,
526
+ "learning_rate": 2.8638071170858157e-05,
527
+ "loss": 0.722,
528
+ "step": 370
529
+ },
530
+ {
531
+ "epoch": 0.5494505494505495,
532
+ "grad_norm": 1.1151386499404907,
533
+ "learning_rate": 2.8587216325115694e-05,
534
+ "loss": 0.7574,
535
+ "step": 375
536
+ },
537
+ {
538
+ "epoch": 0.5567765567765568,
539
+ "grad_norm": 1.1200662851333618,
540
+ "learning_rate": 2.8535476196190807e-05,
541
+ "loss": 0.7204,
542
+ "step": 380
543
+ },
544
+ {
545
+ "epoch": 0.5641025641025641,
546
+ "grad_norm": 1.1838017702102661,
547
+ "learning_rate": 2.848285415524245e-05,
548
+ "loss": 0.7756,
549
+ "step": 385
550
+ },
551
+ {
552
+ "epoch": 0.5714285714285714,
553
+ "grad_norm": 1.1376712322235107,
554
+ "learning_rate": 2.8429353630891085e-05,
555
+ "loss": 0.7384,
556
+ "step": 390
557
+ },
558
+ {
559
+ "epoch": 0.5787545787545788,
560
+ "grad_norm": 1.0719712972640991,
561
+ "learning_rate": 2.837497810899527e-05,
562
+ "loss": 0.764,
563
+ "step": 395
564
+ },
565
+ {
566
+ "epoch": 0.5860805860805861,
567
+ "grad_norm": 1.1748335361480713,
568
+ "learning_rate": 2.8319731132424576e-05,
569
+ "loss": 0.6718,
570
+ "step": 400
571
+ },
572
+ {
573
+ "epoch": 0.5934065934065934,
574
+ "grad_norm": 1.0724573135375977,
575
+ "learning_rate": 2.8263616300828703e-05,
576
+ "loss": 0.7089,
577
+ "step": 405
578
+ },
579
+ {
580
+ "epoch": 0.6007326007326007,
581
+ "grad_norm": 1.1235100030899048,
582
+ "learning_rate": 2.8206637270402963e-05,
583
+ "loss": 0.7111,
584
+ "step": 410
585
+ },
586
+ {
587
+ "epoch": 0.608058608058608,
588
+ "grad_norm": 1.1017916202545166,
589
+ "learning_rate": 2.814879775365007e-05,
590
+ "loss": 0.7187,
591
+ "step": 415
592
+ },
593
+ {
594
+ "epoch": 0.6153846153846154,
595
+ "grad_norm": 1.1182067394256592,
596
+ "learning_rate": 2.809010151913822e-05,
597
+ "loss": 0.6541,
598
+ "step": 420
599
+ },
600
+ {
601
+ "epoch": 0.6227106227106227,
602
+ "grad_norm": 1.2370854616165161,
603
+ "learning_rate": 2.8030552391255592e-05,
604
+ "loss": 0.6732,
605
+ "step": 425
606
+ },
607
+ {
608
+ "epoch": 0.63003663003663,
609
+ "grad_norm": 1.1970367431640625,
610
+ "learning_rate": 2.7970154249961113e-05,
611
+ "loss": 0.675,
612
+ "step": 430
613
+ },
614
+ {
615
+ "epoch": 0.6373626373626373,
616
+ "grad_norm": 1.1454507112503052,
617
+ "learning_rate": 2.79089110305317e-05,
618
+ "loss": 0.6445,
619
+ "step": 435
620
+ },
621
+ {
622
+ "epoch": 0.6446886446886447,
623
+ "grad_norm": 1.1715000867843628,
624
+ "learning_rate": 2.7846826723305842e-05,
625
+ "loss": 0.647,
626
+ "step": 440
627
+ },
628
+ {
629
+ "epoch": 0.652014652014652,
630
+ "grad_norm": 1.2069942951202393,
631
+ "learning_rate": 2.7783905373423593e-05,
632
+ "loss": 0.6969,
633
+ "step": 445
634
+ },
635
+ {
636
+ "epoch": 0.6593406593406593,
637
+ "grad_norm": 1.1830470561981201,
638
+ "learning_rate": 2.772015108056304e-05,
639
+ "loss": 0.6814,
640
+ "step": 450
641
+ },
642
+ {
643
+ "epoch": 0.6666666666666666,
644
+ "grad_norm": 1.2424856424331665,
645
+ "learning_rate": 2.765556799867314e-05,
646
+ "loss": 0.6404,
647
+ "step": 455
648
+ },
649
+ {
650
+ "epoch": 0.673992673992674,
651
+ "grad_norm": 1.345646858215332,
652
+ "learning_rate": 2.7590160335703114e-05,
653
+ "loss": 0.608,
654
+ "step": 460
655
+ },
656
+ {
657
+ "epoch": 0.6813186813186813,
658
+ "grad_norm": 1.099663257598877,
659
+ "learning_rate": 2.752393235332825e-05,
660
+ "loss": 0.6655,
661
+ "step": 465
662
+ },
663
+ {
664
+ "epoch": 0.6886446886446886,
665
+ "grad_norm": 1.3158220052719116,
666
+ "learning_rate": 2.7456888366672233e-05,
667
+ "loss": 0.6398,
668
+ "step": 470
669
+ },
670
+ {
671
+ "epoch": 0.6959706959706959,
672
+ "grad_norm": 1.3298593759536743,
673
+ "learning_rate": 2.7389032744025998e-05,
674
+ "loss": 0.5992,
675
+ "step": 475
676
+ },
677
+ {
678
+ "epoch": 0.7032967032967034,
679
+ "grad_norm": 1.4495782852172852,
680
+ "learning_rate": 2.7320369906563103e-05,
681
+ "loss": 0.5971,
682
+ "step": 480
683
+ },
684
+ {
685
+ "epoch": 0.7106227106227107,
686
+ "grad_norm": 1.19569993019104,
687
+ "learning_rate": 2.7250904328051667e-05,
688
+ "loss": 0.597,
689
+ "step": 485
690
+ },
691
+ {
692
+ "epoch": 0.717948717948718,
693
+ "grad_norm": 1.3111329078674316,
694
+ "learning_rate": 2.7180640534562887e-05,
695
+ "loss": 0.5838,
696
+ "step": 490
697
+ },
698
+ {
699
+ "epoch": 0.7252747252747253,
700
+ "grad_norm": 1.2215338945388794,
701
+ "learning_rate": 2.7109583104176134e-05,
702
+ "loss": 0.5771,
703
+ "step": 495
704
+ },
705
+ {
706
+ "epoch": 0.7326007326007326,
707
+ "grad_norm": 1.3571242094039917,
708
+ "learning_rate": 2.703773666668065e-05,
709
+ "loss": 0.5737,
710
+ "step": 500
711
+ },
712
+ {
713
+ "epoch": 0.73992673992674,
714
+ "grad_norm": 1.4095135927200317,
715
+ "learning_rate": 2.6965105903273923e-05,
716
+ "loss": 0.5867,
717
+ "step": 505
718
+ },
719
+ {
720
+ "epoch": 0.7472527472527473,
721
+ "grad_norm": 1.3398478031158447,
722
+ "learning_rate": 2.6891695546256655e-05,
723
+ "loss": 0.5445,
724
+ "step": 510
725
+ },
726
+ {
727
+ "epoch": 0.7545787545787546,
728
+ "grad_norm": 1.3395978212356567,
729
+ "learning_rate": 2.6817510378724436e-05,
730
+ "loss": 0.5154,
731
+ "step": 515
732
+ },
733
+ {
734
+ "epoch": 0.7619047619047619,
735
+ "grad_norm": 1.2017033100128174,
736
+ "learning_rate": 2.6742555234256093e-05,
737
+ "loss": 0.5638,
738
+ "step": 520
739
+ },
740
+ {
741
+ "epoch": 0.7692307692307693,
742
+ "grad_norm": 1.4696953296661377,
743
+ "learning_rate": 2.6666834996598766e-05,
744
+ "loss": 0.5697,
745
+ "step": 525
746
+ },
747
+ {
748
+ "epoch": 0.7765567765567766,
749
+ "grad_norm": 1.2319661378860474,
750
+ "learning_rate": 2.65903545993497e-05,
751
+ "loss": 0.576,
752
+ "step": 530
753
+ },
754
+ {
755
+ "epoch": 0.7838827838827839,
756
+ "grad_norm": 1.4114476442337036,
757
+ "learning_rate": 2.651311902563478e-05,
758
+ "loss": 0.549,
759
+ "step": 535
760
+ },
761
+ {
762
+ "epoch": 0.7912087912087912,
763
+ "grad_norm": 1.2810288667678833,
764
+ "learning_rate": 2.643513330778388e-05,
765
+ "loss": 0.5135,
766
+ "step": 540
767
+ },
768
+ {
769
+ "epoch": 0.7985347985347986,
770
+ "grad_norm": 1.1400277614593506,
771
+ "learning_rate": 2.635640252700294e-05,
772
+ "loss": 0.5573,
773
+ "step": 545
774
+ },
775
+ {
776
+ "epoch": 0.8058608058608059,
777
+ "grad_norm": 1.4239035844802856,
778
+ "learning_rate": 2.6276931813042942e-05,
779
+ "loss": 0.5112,
780
+ "step": 550
781
+ },
782
+ {
783
+ "epoch": 0.8131868131868132,
784
+ "grad_norm": 1.2470347881317139,
785
+ "learning_rate": 2.619672634386565e-05,
786
+ "loss": 0.5095,
787
+ "step": 555
788
+ },
789
+ {
790
+ "epoch": 0.8205128205128205,
791
+ "grad_norm": 1.110061764717102,
792
+ "learning_rate": 2.6115791345306232e-05,
793
+ "loss": 0.4858,
794
+ "step": 560
795
+ },
796
+ {
797
+ "epoch": 0.8278388278388278,
798
+ "grad_norm": 1.1885781288146973,
799
+ "learning_rate": 2.6034132090732786e-05,
800
+ "loss": 0.5045,
801
+ "step": 565
802
+ },
803
+ {
804
+ "epoch": 0.8351648351648352,
805
+ "grad_norm": 1.2434601783752441,
806
+ "learning_rate": 2.5951753900702743e-05,
807
+ "loss": 0.5416,
808
+ "step": 570
809
+ },
810
+ {
811
+ "epoch": 0.8424908424908425,
812
+ "grad_norm": 1.3534092903137207,
813
+ "learning_rate": 2.58686621426162e-05,
814
+ "loss": 0.5239,
815
+ "step": 575
816
+ },
817
+ {
818
+ "epoch": 0.8498168498168498,
819
+ "grad_norm": 1.1539034843444824,
820
+ "learning_rate": 2.5784862230366206e-05,
821
+ "loss": 0.4937,
822
+ "step": 580
823
+ },
824
+ {
825
+ "epoch": 0.8571428571428571,
826
+ "grad_norm": 1.3080559968948364,
827
+ "learning_rate": 2.570035962398602e-05,
828
+ "loss": 0.5284,
829
+ "step": 585
830
+ },
831
+ {
832
+ "epoch": 0.8644688644688645,
833
+ "grad_norm": 1.4306614398956299,
834
+ "learning_rate": 2.5615159829293356e-05,
835
+ "loss": 0.4945,
836
+ "step": 590
837
+ },
838
+ {
839
+ "epoch": 0.8717948717948718,
840
+ "grad_norm": 1.2547085285186768,
841
+ "learning_rate": 2.5529268397531643e-05,
842
+ "loss": 0.5167,
843
+ "step": 595
844
+ },
845
+ {
846
+ "epoch": 0.8791208791208791,
847
+ "grad_norm": 1.3955950736999512,
848
+ "learning_rate": 2.5442690925008336e-05,
849
+ "loss": 0.5393,
850
+ "step": 600
851
+ },
852
+ {
853
+ "epoch": 0.8864468864468864,
854
+ "grad_norm": 1.1529213190078735,
855
+ "learning_rate": 2.5355433052730296e-05,
856
+ "loss": 0.5081,
857
+ "step": 605
858
+ },
859
+ {
860
+ "epoch": 0.8937728937728938,
861
+ "grad_norm": 1.399857759475708,
862
+ "learning_rate": 2.526750046603622e-05,
863
+ "loss": 0.4698,
864
+ "step": 610
865
+ },
866
+ {
867
+ "epoch": 0.9010989010989011,
868
+ "grad_norm": 1.303623914718628,
869
+ "learning_rate": 2.517889889422624e-05,
870
+ "loss": 0.4526,
871
+ "step": 615
872
+ },
873
+ {
874
+ "epoch": 0.9084249084249084,
875
+ "grad_norm": 1.329766869544983,
876
+ "learning_rate": 2.5089634110188592e-05,
877
+ "loss": 0.469,
878
+ "step": 620
879
+ },
880
+ {
881
+ "epoch": 0.9157509157509157,
882
+ "grad_norm": 1.277982473373413,
883
+ "learning_rate": 2.499971193002353e-05,
884
+ "loss": 0.4714,
885
+ "step": 625
886
+ },
887
+ {
888
+ "epoch": 0.9230769230769231,
889
+ "grad_norm": 1.248856782913208,
890
+ "learning_rate": 2.4909138212664314e-05,
891
+ "loss": 0.4739,
892
+ "step": 630
893
+ },
894
+ {
895
+ "epoch": 0.9304029304029304,
896
+ "grad_norm": 1.21614408493042,
897
+ "learning_rate": 2.4817918859495524e-05,
898
+ "loss": 0.4492,
899
+ "step": 635
900
+ },
901
+ {
902
+ "epoch": 0.9377289377289377,
903
+ "grad_norm": 1.5089861154556274,
904
+ "learning_rate": 2.4726059813968506e-05,
905
+ "loss": 0.4177,
906
+ "step": 640
907
+ },
908
+ {
909
+ "epoch": 0.945054945054945,
910
+ "grad_norm": 1.2757134437561035,
911
+ "learning_rate": 2.4633567061214153e-05,
912
+ "loss": 0.4233,
913
+ "step": 645
914
+ },
915
+ {
916
+ "epoch": 0.9523809523809523,
917
+ "grad_norm": 1.253602385520935,
918
+ "learning_rate": 2.454044662765293e-05,
919
+ "loss": 0.434,
920
+ "step": 650
921
+ },
922
+ {
923
+ "epoch": 0.9597069597069597,
924
+ "grad_norm": 1.2966727018356323,
925
+ "learning_rate": 2.4446704580602222e-05,
926
+ "loss": 0.4717,
927
+ "step": 655
928
+ },
929
+ {
930
+ "epoch": 0.967032967032967,
931
+ "grad_norm": 1.4420007467269897,
932
+ "learning_rate": 2.4352347027881003e-05,
933
+ "loss": 0.4534,
934
+ "step": 660
935
+ },
936
+ {
937
+ "epoch": 0.9743589743589743,
938
+ "grad_norm": 1.1672736406326294,
939
+ "learning_rate": 2.4257380117411893e-05,
940
+ "loss": 0.455,
941
+ "step": 665
942
+ },
943
+ {
944
+ "epoch": 0.9816849816849816,
945
+ "grad_norm": 1.293818712234497,
946
+ "learning_rate": 2.4161810036820577e-05,
947
+ "loss": 0.4418,
948
+ "step": 670
949
+ },
950
+ {
951
+ "epoch": 0.989010989010989,
952
+ "grad_norm": 1.357447862625122,
953
+ "learning_rate": 2.406564301303265e-05,
954
+ "loss": 0.4524,
955
+ "step": 675
956
+ },
957
+ {
958
+ "epoch": 0.9963369963369964,
959
+ "grad_norm": 1.3503096103668213,
960
+ "learning_rate": 2.396888531186789e-05,
961
+ "loss": 0.4231,
962
+ "step": 680
963
+ },
964
+ {
965
+ "epoch": 1.002930402930403,
966
+ "grad_norm": 1.1861598491668701,
967
+ "learning_rate": 2.3871543237632028e-05,
968
+ "loss": 0.3958,
969
+ "step": 685
970
+ },
971
+ {
972
+ "epoch": 1.0102564102564102,
973
+ "grad_norm": 1.3782436847686768,
974
+ "learning_rate": 2.3773623132705955e-05,
975
+ "loss": 0.3487,
976
+ "step": 690
977
+ },
978
+ {
979
+ "epoch": 1.0175824175824175,
980
+ "grad_norm": 1.3742846250534058,
981
+ "learning_rate": 2.3675131377132503e-05,
982
+ "loss": 0.3959,
983
+ "step": 695
984
+ },
985
+ {
986
+ "epoch": 1.0249084249084248,
987
+ "grad_norm": 1.3309235572814941,
988
+ "learning_rate": 2.357607438820074e-05,
989
+ "loss": 0.356,
990
+ "step": 700
991
+ },
992
+ {
993
+ "epoch": 1.0322344322344323,
994
+ "grad_norm": 1.249700903892517,
995
+ "learning_rate": 2.347645862002786e-05,
996
+ "loss": 0.3746,
997
+ "step": 705
998
+ },
999
+ {
1000
+ "epoch": 1.0395604395604396,
1001
+ "grad_norm": 1.238492727279663,
1002
+ "learning_rate": 2.337629056313864e-05,
1003
+ "loss": 0.3243,
1004
+ "step": 710
1005
+ },
1006
+ {
1007
+ "epoch": 1.046886446886447,
1008
+ "grad_norm": 1.3053361177444458,
1009
+ "learning_rate": 2.327557674404256e-05,
1010
+ "loss": 0.3376,
1011
+ "step": 715
1012
+ },
1013
+ {
1014
+ "epoch": 1.0542124542124542,
1015
+ "grad_norm": 1.3403329849243164,
1016
+ "learning_rate": 2.3174323724808565e-05,
1017
+ "loss": 0.3525,
1018
+ "step": 720
1019
+ },
1020
+ {
1021
+ "epoch": 1.0615384615384615,
1022
+ "grad_norm": 1.4044991731643677,
1023
+ "learning_rate": 2.3072538102637507e-05,
1024
+ "loss": 0.3235,
1025
+ "step": 725
1026
+ },
1027
+ {
1028
+ "epoch": 1.0688644688644688,
1029
+ "grad_norm": 1.4254682064056396,
1030
+ "learning_rate": 2.29702265094323e-05,
1031
+ "loss": 0.343,
1032
+ "step": 730
1033
+ },
1034
+ {
1035
+ "epoch": 1.0761904761904761,
1036
+ "grad_norm": 1.240763545036316,
1037
+ "learning_rate": 2.2867395611365807e-05,
1038
+ "loss": 0.3256,
1039
+ "step": 735
1040
+ },
1041
+ {
1042
+ "epoch": 1.0835164835164834,
1043
+ "grad_norm": 1.5764873027801514,
1044
+ "learning_rate": 2.2764052108446523e-05,
1045
+ "loss": 0.3288,
1046
+ "step": 740
1047
+ },
1048
+ {
1049
+ "epoch": 1.0908424908424907,
1050
+ "grad_norm": 1.218450903892517,
1051
+ "learning_rate": 2.2660202734082004e-05,
1052
+ "loss": 0.3411,
1053
+ "step": 745
1054
+ },
1055
+ {
1056
+ "epoch": 1.0981684981684983,
1057
+ "grad_norm": 1.2809207439422607,
1058
+ "learning_rate": 2.2555854254640167e-05,
1059
+ "loss": 0.3123,
1060
+ "step": 750
1061
+ },
1062
+ {
1063
+ "epoch": 1.1054945054945056,
1064
+ "grad_norm": 1.3449431657791138,
1065
+ "learning_rate": 2.2451013469008426e-05,
1066
+ "loss": 0.3191,
1067
+ "step": 755
1068
+ },
1069
+ {
1070
+ "epoch": 1.1128205128205129,
1071
+ "grad_norm": 1.1860313415527344,
1072
+ "learning_rate": 2.2345687208150682e-05,
1073
+ "loss": 0.3221,
1074
+ "step": 760
1075
+ },
1076
+ {
1077
+ "epoch": 1.1201465201465202,
1078
+ "grad_norm": 1.4165265560150146,
1079
+ "learning_rate": 2.2239882334662278e-05,
1080
+ "loss": 0.2941,
1081
+ "step": 765
1082
+ },
1083
+ {
1084
+ "epoch": 1.1274725274725275,
1085
+ "grad_norm": 1.2285124063491821,
1086
+ "learning_rate": 2.2133605742322837e-05,
1087
+ "loss": 0.3005,
1088
+ "step": 770
1089
+ },
1090
+ {
1091
+ "epoch": 1.1347985347985348,
1092
+ "grad_norm": 1.239890456199646,
1093
+ "learning_rate": 2.2026864355647124e-05,
1094
+ "loss": 0.3094,
1095
+ "step": 775
1096
+ },
1097
+ {
1098
+ "epoch": 1.142124542124542,
1099
+ "grad_norm": 1.3588660955429077,
1100
+ "learning_rate": 2.1919665129433828e-05,
1101
+ "loss": 0.3393,
1102
+ "step": 780
1103
+ },
1104
+ {
1105
+ "epoch": 1.1494505494505494,
1106
+ "grad_norm": 1.2373557090759277,
1107
+ "learning_rate": 2.181201504831247e-05,
1108
+ "loss": 0.343,
1109
+ "step": 785
1110
+ },
1111
+ {
1112
+ "epoch": 1.1567765567765567,
1113
+ "grad_norm": 1.3128844499588013,
1114
+ "learning_rate": 2.170392112628828e-05,
1115
+ "loss": 0.3028,
1116
+ "step": 790
1117
+ },
1118
+ {
1119
+ "epoch": 1.1641025641025642,
1120
+ "grad_norm": 1.1008199453353882,
1121
+ "learning_rate": 2.1595390406285214e-05,
1122
+ "loss": 0.2834,
1123
+ "step": 795
1124
+ },
1125
+ {
1126
+ "epoch": 1.1714285714285715,
1127
+ "grad_norm": 1.2620314359664917,
1128
+ "learning_rate": 2.148642995968705e-05,
1129
+ "loss": 0.3097,
1130
+ "step": 800
1131
+ },
1132
+ {
1133
+ "epoch": 1.1787545787545788,
1134
+ "grad_norm": 1.391574740409851,
1135
+ "learning_rate": 2.1377046885876667e-05,
1136
+ "loss": 0.3108,
1137
+ "step": 805
1138
+ },
1139
+ {
1140
+ "epoch": 1.186080586080586,
1141
+ "grad_norm": 1.2171621322631836,
1142
+ "learning_rate": 2.1267248311773468e-05,
1143
+ "loss": 0.3223,
1144
+ "step": 810
1145
+ },
1146
+ {
1147
+ "epoch": 1.1934065934065934,
1148
+ "grad_norm": 1.2164249420166016,
1149
+ "learning_rate": 2.1157041391369035e-05,
1150
+ "loss": 0.2865,
1151
+ "step": 815
1152
+ },
1153
+ {
1154
+ "epoch": 1.2007326007326007,
1155
+ "grad_norm": 1.4462814331054688,
1156
+ "learning_rate": 2.1046433305260983e-05,
1157
+ "loss": 0.2904,
1158
+ "step": 820
1159
+ },
1160
+ {
1161
+ "epoch": 1.208058608058608,
1162
+ "grad_norm": 1.2449432611465454,
1163
+ "learning_rate": 2.0935431260185124e-05,
1164
+ "loss": 0.2554,
1165
+ "step": 825
1166
+ },
1167
+ {
1168
+ "epoch": 1.2153846153846155,
1169
+ "grad_norm": 1.1743241548538208,
1170
+ "learning_rate": 2.0824042488545904e-05,
1171
+ "loss": 0.2948,
1172
+ "step": 830
1173
+ },
1174
+ {
1175
+ "epoch": 1.2227106227106228,
1176
+ "grad_norm": 1.2620198726654053,
1177
+ "learning_rate": 2.0712274247945162e-05,
1178
+ "loss": 0.3103,
1179
+ "step": 835
1180
+ },
1181
+ {
1182
+ "epoch": 1.23003663003663,
1183
+ "grad_norm": 1.3025723695755005,
1184
+ "learning_rate": 2.0600133820709264e-05,
1185
+ "loss": 0.2885,
1186
+ "step": 840
1187
+ },
1188
+ {
1189
+ "epoch": 1.2373626373626374,
1190
+ "grad_norm": 1.338291883468628,
1191
+ "learning_rate": 2.048762851341463e-05,
1192
+ "loss": 0.2761,
1193
+ "step": 845
1194
+ },
1195
+ {
1196
+ "epoch": 1.2446886446886447,
1197
+ "grad_norm": 1.2580403089523315,
1198
+ "learning_rate": 2.0374765656411653e-05,
1199
+ "loss": 0.2811,
1200
+ "step": 850
1201
+ },
1202
+ {
1203
+ "epoch": 1.252014652014652,
1204
+ "grad_norm": 1.262098789215088,
1205
+ "learning_rate": 2.0261552603347075e-05,
1206
+ "loss": 0.2928,
1207
+ "step": 855
1208
+ },
1209
+ {
1210
+ "epoch": 1.2593406593406593,
1211
+ "grad_norm": 1.2450367212295532,
1212
+ "learning_rate": 2.0147996730684888e-05,
1213
+ "loss": 0.2891,
1214
+ "step": 860
1215
+ },
1216
+ {
1217
+ "epoch": 1.2666666666666666,
1218
+ "grad_norm": 1.292145013809204,
1219
+ "learning_rate": 2.00341054372257e-05,
1220
+ "loss": 0.273,
1221
+ "step": 865
1222
+ },
1223
+ {
1224
+ "epoch": 1.273992673992674,
1225
+ "grad_norm": 1.316749095916748,
1226
+ "learning_rate": 1.9919886143624652e-05,
1227
+ "loss": 0.2702,
1228
+ "step": 870
1229
+ },
1230
+ {
1231
+ "epoch": 1.2813186813186812,
1232
+ "grad_norm": 1.1842753887176514,
1233
+ "learning_rate": 1.9805346291907927e-05,
1234
+ "loss": 0.2759,
1235
+ "step": 875
1236
+ },
1237
+ {
1238
+ "epoch": 1.2886446886446885,
1239
+ "grad_norm": 1.2208205461502075,
1240
+ "learning_rate": 1.9690493344987878e-05,
1241
+ "loss": 0.2799,
1242
+ "step": 880
1243
+ },
1244
+ {
1245
+ "epoch": 1.295970695970696,
1246
+ "grad_norm": 1.3807168006896973,
1247
+ "learning_rate": 1.9575334786176752e-05,
1248
+ "loss": 0.2644,
1249
+ "step": 885
1250
+ },
1251
+ {
1252
+ "epoch": 1.3032967032967033,
1253
+ "grad_norm": 1.2988091707229614,
1254
+ "learning_rate": 1.9459878118699124e-05,
1255
+ "loss": 0.2457,
1256
+ "step": 890
1257
+ },
1258
+ {
1259
+ "epoch": 1.3106227106227106,
1260
+ "grad_norm": 1.1406099796295166,
1261
+ "learning_rate": 1.934413086520301e-05,
1262
+ "loss": 0.28,
1263
+ "step": 895
1264
+ },
1265
+ {
1266
+ "epoch": 1.317948717948718,
1267
+ "grad_norm": 1.1036453247070312,
1268
+ "learning_rate": 1.9228100567269745e-05,
1269
+ "loss": 0.2759,
1270
+ "step": 900
1271
+ },
1272
+ {
1273
+ "epoch": 1.3252747252747252,
1274
+ "grad_norm": 1.329750895500183,
1275
+ "learning_rate": 1.9111794784922592e-05,
1276
+ "loss": 0.2862,
1277
+ "step": 905
1278
+ },
1279
+ {
1280
+ "epoch": 1.3326007326007325,
1281
+ "grad_norm": 1.0778213739395142,
1282
+ "learning_rate": 1.899522109613414e-05,
1283
+ "loss": 0.2406,
1284
+ "step": 910
1285
+ },
1286
+ {
1287
+ "epoch": 1.33992673992674,
1288
+ "grad_norm": 1.267181396484375,
1289
+ "learning_rate": 1.8878387096332616e-05,
1290
+ "loss": 0.252,
1291
+ "step": 915
1292
+ },
1293
+ {
1294
+ "epoch": 1.3472527472527474,
1295
+ "grad_norm": 1.1901800632476807,
1296
+ "learning_rate": 1.876130039790695e-05,
1297
+ "loss": 0.2295,
1298
+ "step": 920
1299
+ },
1300
+ {
1301
+ "epoch": 1.3545787545787547,
1302
+ "grad_norm": 1.1847642660140991,
1303
+ "learning_rate": 1.8643968629710797e-05,
1304
+ "loss": 0.2299,
1305
+ "step": 925
1306
+ },
1307
+ {
1308
+ "epoch": 1.361904761904762,
1309
+ "grad_norm": 1.0492013692855835,
1310
+ "learning_rate": 1.8526399436565484e-05,
1311
+ "loss": 0.2435,
1312
+ "step": 930
1313
+ },
1314
+ {
1315
+ "epoch": 1.3692307692307693,
1316
+ "grad_norm": 1.1790529489517212,
1317
+ "learning_rate": 1.840860047876191e-05,
1318
+ "loss": 0.2642,
1319
+ "step": 935
1320
+ },
1321
+ {
1322
+ "epoch": 1.3765567765567766,
1323
+ "grad_norm": 1.3635669946670532,
1324
+ "learning_rate": 1.8290579431561413e-05,
1325
+ "loss": 0.2281,
1326
+ "step": 940
1327
+ },
1328
+ {
1329
+ "epoch": 1.3838827838827839,
1330
+ "grad_norm": 1.1547739505767822,
1331
+ "learning_rate": 1.8172343984695716e-05,
1332
+ "loss": 0.2085,
1333
+ "step": 945
1334
+ },
1335
+ {
1336
+ "epoch": 1.3912087912087912,
1337
+ "grad_norm": 1.1450775861740112,
1338
+ "learning_rate": 1.8053901841865856e-05,
1339
+ "loss": 0.2297,
1340
+ "step": 950
1341
+ },
1342
+ {
1343
+ "epoch": 1.3985347985347985,
1344
+ "grad_norm": 1.3273813724517822,
1345
+ "learning_rate": 1.7935260720240292e-05,
1346
+ "loss": 0.2626,
1347
+ "step": 955
1348
+ },
1349
+ {
1350
+ "epoch": 1.4058608058608058,
1351
+ "grad_norm": 1.2674297094345093,
1352
+ "learning_rate": 1.7816428349952056e-05,
1353
+ "loss": 0.2409,
1354
+ "step": 960
1355
+ },
1356
+ {
1357
+ "epoch": 1.413186813186813,
1358
+ "grad_norm": 1.315129280090332,
1359
+ "learning_rate": 1.7697412473595088e-05,
1360
+ "loss": 0.2397,
1361
+ "step": 965
1362
+ },
1363
+ {
1364
+ "epoch": 1.4205128205128206,
1365
+ "grad_norm": 1.2142728567123413,
1366
+ "learning_rate": 1.75782208457198e-05,
1367
+ "loss": 0.2292,
1368
+ "step": 970
1369
+ },
1370
+ {
1371
+ "epoch": 1.4278388278388279,
1372
+ "grad_norm": 1.3023855686187744,
1373
+ "learning_rate": 1.7458861232327785e-05,
1374
+ "loss": 0.2521,
1375
+ "step": 975
1376
+ },
1377
+ {
1378
+ "epoch": 1.4351648351648352,
1379
+ "grad_norm": 1.1531919240951538,
1380
+ "learning_rate": 1.733934141036583e-05,
1381
+ "loss": 0.234,
1382
+ "step": 980
1383
+ },
1384
+ {
1385
+ "epoch": 1.4424908424908425,
1386
+ "grad_norm": 1.3371647596359253,
1387
+ "learning_rate": 1.721966916721921e-05,
1388
+ "loss": 0.2292,
1389
+ "step": 985
1390
+ },
1391
+ {
1392
+ "epoch": 1.4498168498168498,
1393
+ "grad_norm": 1.1629012823104858,
1394
+ "learning_rate": 1.7099852300204296e-05,
1395
+ "loss": 0.224,
1396
+ "step": 990
1397
+ },
1398
+ {
1399
+ "epoch": 1.457142857142857,
1400
+ "grad_norm": 1.2537931203842163,
1401
+ "learning_rate": 1.6979898616060514e-05,
1402
+ "loss": 0.2016,
1403
+ "step": 995
1404
+ },
1405
+ {
1406
+ "epoch": 1.4644688644688646,
1407
+ "grad_norm": 1.4359750747680664,
1408
+ "learning_rate": 1.6859815930441683e-05,
1409
+ "loss": 0.212,
1410
+ "step": 1000
1411
+ },
1412
+ {
1413
+ "epoch": 1.471794871794872,
1414
+ "grad_norm": 1.0480167865753174,
1415
+ "learning_rate": 1.6739612067406798e-05,
1416
+ "loss": 0.1865,
1417
+ "step": 1005
1418
+ },
1419
+ {
1420
+ "epoch": 1.4791208791208792,
1421
+ "grad_norm": 1.29014253616333,
1422
+ "learning_rate": 1.6619294858910242e-05,
1423
+ "loss": 0.2171,
1424
+ "step": 1010
1425
+ },
1426
+ {
1427
+ "epoch": 1.4864468864468865,
1428
+ "grad_norm": 1.3352724313735962,
1429
+ "learning_rate": 1.649887214429149e-05,
1430
+ "loss": 0.209,
1431
+ "step": 1015
1432
+ },
1433
+ {
1434
+ "epoch": 1.4937728937728938,
1435
+ "grad_norm": 1.067258596420288,
1436
+ "learning_rate": 1.6378351769764316e-05,
1437
+ "loss": 0.192,
1438
+ "step": 1020
1439
+ },
1440
+ {
1441
+ "epoch": 1.501098901098901,
1442
+ "grad_norm": 1.2493972778320312,
1443
+ "learning_rate": 1.6257741587905596e-05,
1444
+ "loss": 0.2099,
1445
+ "step": 1025
1446
+ },
1447
+ {
1448
+ "epoch": 1.5084249084249084,
1449
+ "grad_norm": 1.2061505317687988,
1450
+ "learning_rate": 1.6137049457143667e-05,
1451
+ "loss": 0.2045,
1452
+ "step": 1030
1453
+ },
1454
+ {
1455
+ "epoch": 1.5157509157509157,
1456
+ "grad_norm": 1.2795240879058838,
1457
+ "learning_rate": 1.6016283241246274e-05,
1458
+ "loss": 0.2073,
1459
+ "step": 1035
1460
+ },
1461
+ {
1462
+ "epoch": 1.523076923076923,
1463
+ "grad_norm": 1.1073862314224243,
1464
+ "learning_rate": 1.589545080880823e-05,
1465
+ "loss": 0.1987,
1466
+ "step": 1040
1467
+ },
1468
+ {
1469
+ "epoch": 1.5304029304029303,
1470
+ "grad_norm": 1.2649872303009033,
1471
+ "learning_rate": 1.577456003273874e-05,
1472
+ "loss": 0.2089,
1473
+ "step": 1045
1474
+ },
1475
+ {
1476
+ "epoch": 1.5377289377289376,
1477
+ "grad_norm": 1.4907593727111816,
1478
+ "learning_rate": 1.5653618789748405e-05,
1479
+ "loss": 0.2163,
1480
+ "step": 1050
1481
+ },
1482
+ {
1483
+ "epoch": 1.545054945054945,
1484
+ "grad_norm": 1.160706877708435,
1485
+ "learning_rate": 1.553263495983604e-05,
1486
+ "loss": 0.1754,
1487
+ "step": 1055
1488
+ },
1489
+ {
1490
+ "epoch": 1.5523809523809524,
1491
+ "grad_norm": 1.4039742946624756,
1492
+ "learning_rate": 1.5411616425775223e-05,
1493
+ "loss": 0.2363,
1494
+ "step": 1060
1495
+ },
1496
+ {
1497
+ "epoch": 1.5597069597069597,
1498
+ "grad_norm": 1.232223391532898,
1499
+ "learning_rate": 1.5290571072600718e-05,
1500
+ "loss": 0.2245,
1501
+ "step": 1065
1502
+ },
1503
+ {
1504
+ "epoch": 1.567032967032967,
1505
+ "grad_norm": 1.131118655204773,
1506
+ "learning_rate": 1.5169506787094698e-05,
1507
+ "loss": 0.183,
1508
+ "step": 1070
1509
+ },
1510
+ {
1511
+ "epoch": 1.5743589743589743,
1512
+ "grad_norm": 1.162467360496521,
1513
+ "learning_rate": 1.5048431457272877e-05,
1514
+ "loss": 0.1764,
1515
+ "step": 1075
1516
+ },
1517
+ {
1518
+ "epoch": 1.5816849816849818,
1519
+ "grad_norm": 1.2327086925506592,
1520
+ "learning_rate": 1.4927352971870576e-05,
1521
+ "loss": 0.2027,
1522
+ "step": 1080
1523
+ },
1524
+ {
1525
+ "epoch": 1.5890109890109891,
1526
+ "grad_norm": 1.1661217212677002,
1527
+ "learning_rate": 1.4806279219828713e-05,
1528
+ "loss": 0.2159,
1529
+ "step": 1085
1530
+ },
1531
+ {
1532
+ "epoch": 1.5963369963369964,
1533
+ "grad_norm": 1.1822779178619385,
1534
+ "learning_rate": 1.4685218089779813e-05,
1535
+ "loss": 0.1972,
1536
+ "step": 1090
1537
+ },
1538
+ {
1539
+ "epoch": 1.6036630036630037,
1540
+ "grad_norm": 1.2681963443756104,
1541
+ "learning_rate": 1.4564177469533994e-05,
1542
+ "loss": 0.1811,
1543
+ "step": 1095
1544
+ },
1545
+ {
1546
+ "epoch": 1.610989010989011,
1547
+ "grad_norm": 1.400927186012268,
1548
+ "learning_rate": 1.4443165245565046e-05,
1549
+ "loss": 0.1738,
1550
+ "step": 1100
1551
+ },
1552
+ {
1553
+ "epoch": 1.6183150183150183,
1554
+ "grad_norm": 1.1321874856948853,
1555
+ "learning_rate": 1.4322189302496597e-05,
1556
+ "loss": 0.2033,
1557
+ "step": 1105
1558
+ },
1559
+ {
1560
+ "epoch": 1.6256410256410256,
1561
+ "grad_norm": 1.1278076171875,
1562
+ "learning_rate": 1.4201257522588352e-05,
1563
+ "loss": 0.1629,
1564
+ "step": 1110
1565
+ },
1566
+ {
1567
+ "epoch": 1.632967032967033,
1568
+ "grad_norm": 1.1997780799865723,
1569
+ "learning_rate": 1.4080377785222563e-05,
1570
+ "loss": 0.179,
1571
+ "step": 1115
1572
+ },
1573
+ {
1574
+ "epoch": 1.6402930402930402,
1575
+ "grad_norm": 1.1380888223648071,
1576
+ "learning_rate": 1.395955796639059e-05,
1577
+ "loss": 0.1716,
1578
+ "step": 1120
1579
+ },
1580
+ {
1581
+ "epoch": 1.6476190476190475,
1582
+ "grad_norm": 0.9687121510505676,
1583
+ "learning_rate": 1.3838805938179788e-05,
1584
+ "loss": 0.1756,
1585
+ "step": 1125
1586
+ },
1587
+ {
1588
+ "epoch": 1.6549450549450548,
1589
+ "grad_norm": 1.0535095930099487,
1590
+ "learning_rate": 1.3718129568260568e-05,
1591
+ "loss": 0.1735,
1592
+ "step": 1130
1593
+ },
1594
+ {
1595
+ "epoch": 1.6622710622710621,
1596
+ "grad_norm": 1.1114075183868408,
1597
+ "learning_rate": 1.3597536719373767e-05,
1598
+ "loss": 0.1706,
1599
+ "step": 1135
1600
+ },
1601
+ {
1602
+ "epoch": 1.6695970695970694,
1603
+ "grad_norm": 1.0749149322509766,
1604
+ "learning_rate": 1.347703524881838e-05,
1605
+ "loss": 0.1792,
1606
+ "step": 1140
1607
+ },
1608
+ {
1609
+ "epoch": 1.676923076923077,
1610
+ "grad_norm": 1.2217072248458862,
1611
+ "learning_rate": 1.3356633007939577e-05,
1612
+ "loss": 0.1654,
1613
+ "step": 1145
1614
+ },
1615
+ {
1616
+ "epoch": 1.6842490842490843,
1617
+ "grad_norm": 1.344526767730713,
1618
+ "learning_rate": 1.3236337841617173e-05,
1619
+ "loss": 0.1777,
1620
+ "step": 1150
1621
+ },
1622
+ {
1623
+ "epoch": 1.6915750915750916,
1624
+ "grad_norm": 1.0998477935791016,
1625
+ "learning_rate": 1.3116157587754455e-05,
1626
+ "loss": 0.1538,
1627
+ "step": 1155
1628
+ },
1629
+ {
1630
+ "epoch": 1.6989010989010989,
1631
+ "grad_norm": 1.046402931213379,
1632
+ "learning_rate": 1.2996100076767544e-05,
1633
+ "loss": 0.1602,
1634
+ "step": 1160
1635
+ },
1636
+ {
1637
+ "epoch": 1.7062271062271064,
1638
+ "grad_norm": 1.0636180639266968,
1639
+ "learning_rate": 1.2876173131075164e-05,
1640
+ "loss": 0.1901,
1641
+ "step": 1165
1642
+ },
1643
+ {
1644
+ "epoch": 1.7135531135531137,
1645
+ "grad_norm": 1.1561256647109985,
1646
+ "learning_rate": 1.2756384564588969e-05,
1647
+ "loss": 0.1697,
1648
+ "step": 1170
1649
+ },
1650
+ {
1651
+ "epoch": 1.720879120879121,
1652
+ "grad_norm": 1.0080074071884155,
1653
+ "learning_rate": 1.2636742182204453e-05,
1654
+ "loss": 0.1625,
1655
+ "step": 1175
1656
+ },
1657
+ {
1658
+ "epoch": 1.7282051282051283,
1659
+ "grad_norm": 0.9807740449905396,
1660
+ "learning_rate": 1.2517253779292393e-05,
1661
+ "loss": 0.1647,
1662
+ "step": 1180
1663
+ },
1664
+ {
1665
+ "epoch": 1.7355311355311356,
1666
+ "grad_norm": 1.1684633493423462,
1667
+ "learning_rate": 1.2397927141190937e-05,
1668
+ "loss": 0.1492,
1669
+ "step": 1185
1670
+ },
1671
+ {
1672
+ "epoch": 1.7428571428571429,
1673
+ "grad_norm": 1.0675970315933228,
1674
+ "learning_rate": 1.227877004269834e-05,
1675
+ "loss": 0.1624,
1676
+ "step": 1190
1677
+ },
1678
+ {
1679
+ "epoch": 1.7501831501831502,
1680
+ "grad_norm": 1.0883363485336304,
1681
+ "learning_rate": 1.2159790247566425e-05,
1682
+ "loss": 0.169,
1683
+ "step": 1195
1684
+ },
1685
+ {
1686
+ "epoch": 1.7575091575091575,
1687
+ "grad_norm": 1.00874924659729,
1688
+ "learning_rate": 1.2040995507994698e-05,
1689
+ "loss": 0.1479,
1690
+ "step": 1200
1691
+ },
1692
+ {
1693
+ "epoch": 1.7648351648351648,
1694
+ "grad_norm": 1.154402494430542,
1695
+ "learning_rate": 1.192239356412525e-05,
1696
+ "loss": 0.1612,
1697
+ "step": 1205
1698
+ },
1699
+ {
1700
+ "epoch": 1.772161172161172,
1701
+ "grad_norm": 1.1723260879516602,
1702
+ "learning_rate": 1.1803992143538468e-05,
1703
+ "loss": 0.1652,
1704
+ "step": 1210
1705
+ },
1706
+ {
1707
+ "epoch": 1.7794871794871794,
1708
+ "grad_norm": 1.0744177103042603,
1709
+ "learning_rate": 1.168579896074952e-05,
1710
+ "loss": 0.1616,
1711
+ "step": 1215
1712
+ },
1713
+ {
1714
+ "epoch": 1.7868131868131867,
1715
+ "grad_norm": 1.3360925912857056,
1716
+ "learning_rate": 1.156782171670572e-05,
1717
+ "loss": 0.1701,
1718
+ "step": 1220
1719
+ },
1720
+ {
1721
+ "epoch": 1.794139194139194,
1722
+ "grad_norm": 1.158911943435669,
1723
+ "learning_rate": 1.1450068098284748e-05,
1724
+ "loss": 0.1586,
1725
+ "step": 1225
1726
+ },
1727
+ {
1728
+ "epoch": 1.8014652014652015,
1729
+ "grad_norm": 1.1165605783462524,
1730
+ "learning_rate": 1.133254577779384e-05,
1731
+ "loss": 0.1572,
1732
+ "step": 1230
1733
+ },
1734
+ {
1735
+ "epoch": 1.8087912087912088,
1736
+ "grad_norm": 1.0566843748092651,
1737
+ "learning_rate": 1.121526241246988e-05,
1738
+ "loss": 0.1215,
1739
+ "step": 1235
1740
+ },
1741
+ {
1742
+ "epoch": 1.816117216117216,
1743
+ "grad_norm": 1.3197264671325684,
1744
+ "learning_rate": 1.1098225643980468e-05,
1745
+ "loss": 0.1465,
1746
+ "step": 1240
1747
+ },
1748
+ {
1749
+ "epoch": 1.8234432234432234,
1750
+ "grad_norm": 0.9467335939407349,
1751
+ "learning_rate": 1.0981443097926047e-05,
1752
+ "loss": 0.1531,
1753
+ "step": 1245
1754
+ },
1755
+ {
1756
+ "epoch": 1.830769230769231,
1757
+ "grad_norm": 1.0139106512069702,
1758
+ "learning_rate": 1.0864922383343061e-05,
1759
+ "loss": 0.1403,
1760
+ "step": 1250
1761
+ },
1762
+ {
1763
+ "epoch": 1.8380952380952382,
1764
+ "grad_norm": 1.0302916765213013,
1765
+ "learning_rate": 1.0748671092208154e-05,
1766
+ "loss": 0.136,
1767
+ "step": 1255
1768
+ },
1769
+ {
1770
+ "epoch": 1.8454212454212455,
1771
+ "grad_norm": 1.093135952949524,
1772
+ "learning_rate": 1.0632696798943514e-05,
1773
+ "loss": 0.137,
1774
+ "step": 1260
1775
+ },
1776
+ {
1777
+ "epoch": 1.8527472527472528,
1778
+ "grad_norm": 0.9857581257820129,
1779
+ "learning_rate": 1.0517007059923387e-05,
1780
+ "loss": 0.1332,
1781
+ "step": 1265
1782
+ },
1783
+ {
1784
+ "epoch": 1.8600732600732601,
1785
+ "grad_norm": 0.9491907358169556,
1786
+ "learning_rate": 1.040160941298172e-05,
1787
+ "loss": 0.1447,
1788
+ "step": 1270
1789
+ },
1790
+ {
1791
+ "epoch": 1.8673992673992674,
1792
+ "grad_norm": 1.127459168434143,
1793
+ "learning_rate": 1.0286511376921007e-05,
1794
+ "loss": 0.132,
1795
+ "step": 1275
1796
+ },
1797
+ {
1798
+ "epoch": 1.8747252747252747,
1799
+ "grad_norm": 1.0860737562179565,
1800
+ "learning_rate": 1.0171720451022424e-05,
1801
+ "loss": 0.1292,
1802
+ "step": 1280
1803
+ },
1804
+ {
1805
+ "epoch": 1.882051282051282,
1806
+ "grad_norm": 1.0091402530670166,
1807
+ "learning_rate": 1.0057244114557213e-05,
1808
+ "loss": 0.1461,
1809
+ "step": 1285
1810
+ },
1811
+ {
1812
+ "epoch": 1.8893772893772893,
1813
+ "grad_norm": 0.9668272137641907,
1814
+ "learning_rate": 9.94308982629934e-06,
1815
+ "loss": 0.1424,
1816
+ "step": 1290
1817
+ },
1818
+ {
1819
+ "epoch": 1.8967032967032966,
1820
+ "grad_norm": 0.9310147762298584,
1821
+ "learning_rate": 9.82926502403953e-06,
1822
+ "loss": 0.1213,
1823
+ "step": 1295
1824
+ },
1825
+ {
1826
+ "epoch": 1.904029304029304,
1827
+ "grad_norm": 1.0682415962219238,
1828
+ "learning_rate": 9.715777124100657e-06,
1829
+ "loss": 0.1325,
1830
+ "step": 1300
1831
+ },
1832
+ {
1833
+ "epoch": 1.9113553113553112,
1834
+ "grad_norm": 1.2194106578826904,
1835
+ "learning_rate": 9.602633520854522e-06,
1836
+ "loss": 0.1318,
1837
+ "step": 1305
1838
+ },
1839
+ {
1840
+ "epoch": 1.9186813186813185,
1841
+ "grad_norm": 0.9982204437255859,
1842
+ "learning_rate": 9.489841586240065e-06,
1843
+ "loss": 0.1226,
1844
+ "step": 1310
1845
+ },
1846
+ {
1847
+ "epoch": 1.926007326007326,
1848
+ "grad_norm": 1.1038477420806885,
1849
+ "learning_rate": 9.377408669283047e-06,
1850
+ "loss": 0.1385,
1851
+ "step": 1315
1852
+ },
1853
+ {
1854
+ "epoch": 1.9333333333333333,
1855
+ "grad_norm": 0.9194656014442444,
1856
+ "learning_rate": 9.265342095617224e-06,
1857
+ "loss": 0.1217,
1858
+ "step": 1320
1859
+ },
1860
+ {
1861
+ "epoch": 1.9406593406593406,
1862
+ "grad_norm": 0.9665452837944031,
1863
+ "learning_rate": 9.15364916700704e-06,
1864
+ "loss": 0.127,
1865
+ "step": 1325
1866
+ },
1867
+ {
1868
+ "epoch": 1.947985347985348,
1869
+ "grad_norm": 0.9396631121635437,
1870
+ "learning_rate": 9.042337160871861e-06,
1871
+ "loss": 0.131,
1872
+ "step": 1330
1873
+ },
1874
+ {
1875
+ "epoch": 1.9553113553113555,
1876
+ "grad_norm": 0.9682226181030273,
1877
+ "learning_rate": 8.931413329811827e-06,
1878
+ "loss": 0.1208,
1879
+ "step": 1335
1880
+ },
1881
+ {
1882
+ "epoch": 1.9626373626373628,
1883
+ "grad_norm": 1.3224050998687744,
1884
+ "learning_rate": 8.820884901135308e-06,
1885
+ "loss": 0.1377,
1886
+ "step": 1340
1887
+ },
1888
+ {
1889
+ "epoch": 1.96996336996337,
1890
+ "grad_norm": 1.1851681470870972,
1891
+ "learning_rate": 8.710759076387977e-06,
1892
+ "loss": 0.1285,
1893
+ "step": 1345
1894
+ },
1895
+ {
1896
+ "epoch": 1.9772893772893774,
1897
+ "grad_norm": 0.997458815574646,
1898
+ "learning_rate": 8.601043030883626e-06,
1899
+ "loss": 0.1216,
1900
+ "step": 1350
1901
+ },
1902
+ {
1903
+ "epoch": 1.9846153846153847,
1904
+ "grad_norm": 1.0154951810836792,
1905
+ "learning_rate": 8.491743913236629e-06,
1906
+ "loss": 0.1304,
1907
+ "step": 1355
1908
+ },
1909
+ {
1910
+ "epoch": 1.991941391941392,
1911
+ "grad_norm": 0.9404909610748291,
1912
+ "learning_rate": 8.382868844896176e-06,
1913
+ "loss": 0.1366,
1914
+ "step": 1360
1915
+ },
1916
+ {
1917
+ "epoch": 1.9992673992673993,
1918
+ "grad_norm": 0.9820082783699036,
1919
+ "learning_rate": 8.274424919682274e-06,
1920
+ "loss": 0.1362,
1921
+ "step": 1365
1922
+ },
1923
+ {
1924
+ "epoch": 2.005860805860806,
1925
+ "grad_norm": 1.0059517621994019,
1926
+ "learning_rate": 8.166419203323538e-06,
1927
+ "loss": 0.0945,
1928
+ "step": 1370
1929
+ },
1930
+ {
1931
+ "epoch": 2.013186813186813,
1932
+ "grad_norm": 0.8542543053627014,
1933
+ "learning_rate": 8.058858732996829e-06,
1934
+ "loss": 0.0859,
1935
+ "step": 1375
1936
+ },
1937
+ {
1938
+ "epoch": 2.0205128205128204,
1939
+ "grad_norm": 1.002951979637146,
1940
+ "learning_rate": 7.951750516868733e-06,
1941
+ "loss": 0.0976,
1942
+ "step": 1380
1943
+ },
1944
+ {
1945
+ "epoch": 2.0278388278388277,
1946
+ "grad_norm": 0.8994050621986389,
1947
+ "learning_rate": 7.845101533638944e-06,
1948
+ "loss": 0.0891,
1949
+ "step": 1385
1950
+ },
1951
+ {
1952
+ "epoch": 2.035164835164835,
1953
+ "grad_norm": 0.8844568729400635,
1954
+ "learning_rate": 7.738918732085573e-06,
1955
+ "loss": 0.0967,
1956
+ "step": 1390
1957
+ },
1958
+ {
1959
+ "epoch": 2.0424908424908423,
1960
+ "grad_norm": 0.7864266633987427,
1961
+ "learning_rate": 7.633209030612379e-06,
1962
+ "loss": 0.0984,
1963
+ "step": 1395
1964
+ },
1965
+ {
1966
+ "epoch": 2.0498168498168496,
1967
+ "grad_norm": 0.747887909412384,
1968
+ "learning_rate": 7.527979316797983e-06,
1969
+ "loss": 0.0959,
1970
+ "step": 1400
1971
+ },
1972
+ {
1973
+ "epoch": 2.057142857142857,
1974
+ "grad_norm": 0.8532736897468567,
1975
+ "learning_rate": 7.4232364469471605e-06,
1976
+ "loss": 0.0954,
1977
+ "step": 1405
1978
+ },
1979
+ {
1980
+ "epoch": 2.0644688644688647,
1981
+ "grad_norm": 1.0058013200759888,
1982
+ "learning_rate": 7.318987245644056e-06,
1983
+ "loss": 0.1002,
1984
+ "step": 1410
1985
+ },
1986
+ {
1987
+ "epoch": 2.071794871794872,
1988
+ "grad_norm": 0.9096917510032654,
1989
+ "learning_rate": 7.215238505307547e-06,
1990
+ "loss": 0.1014,
1991
+ "step": 1415
1992
+ },
1993
+ {
1994
+ "epoch": 2.0791208791208793,
1995
+ "grad_norm": 0.7891750335693359,
1996
+ "learning_rate": 7.1119969857486825e-06,
1997
+ "loss": 0.1115,
1998
+ "step": 1420
1999
+ },
2000
+ {
2001
+ "epoch": 2.0864468864468866,
2002
+ "grad_norm": 0.9288976788520813,
2003
+ "learning_rate": 7.009269413730236e-06,
2004
+ "loss": 0.087,
2005
+ "step": 1425
2006
+ },
2007
+ {
2008
+ "epoch": 2.093772893772894,
2009
+ "grad_norm": 0.9009776711463928,
2010
+ "learning_rate": 6.907062482528427e-06,
2011
+ "loss": 0.0981,
2012
+ "step": 1430
2013
+ },
2014
+ {
2015
+ "epoch": 2.101098901098901,
2016
+ "grad_norm": 0.7330975532531738,
2017
+ "learning_rate": 6.805382851496807e-06,
2018
+ "loss": 0.0962,
2019
+ "step": 1435
2020
+ },
2021
+ {
2022
+ "epoch": 2.1084249084249085,
2023
+ "grad_norm": 0.7801343202590942,
2024
+ "learning_rate": 6.704237145632369e-06,
2025
+ "loss": 0.0965,
2026
+ "step": 1440
2027
+ },
2028
+ {
2029
+ "epoch": 2.115750915750916,
2030
+ "grad_norm": 0.939649760723114,
2031
+ "learning_rate": 6.603631955143911e-06,
2032
+ "loss": 0.0982,
2033
+ "step": 1445
2034
+ },
2035
+ {
2036
+ "epoch": 2.123076923076923,
2037
+ "grad_norm": 0.8976964354515076,
2038
+ "learning_rate": 6.503573835022601e-06,
2039
+ "loss": 0.0895,
2040
+ "step": 1450
2041
+ },
2042
+ {
2043
+ "epoch": 2.1304029304029304,
2044
+ "grad_norm": 1.068953514099121,
2045
+ "learning_rate": 6.404069304614925e-06,
2046
+ "loss": 0.0907,
2047
+ "step": 1455
2048
+ },
2049
+ {
2050
+ "epoch": 2.1377289377289377,
2051
+ "grad_norm": 0.9661380648612976,
2052
+ "learning_rate": 6.305124847197912e-06,
2053
+ "loss": 0.0942,
2054
+ "step": 1460
2055
+ },
2056
+ {
2057
+ "epoch": 2.145054945054945,
2058
+ "grad_norm": 0.8189408779144287,
2059
+ "learning_rate": 6.2067469095566896e-06,
2060
+ "loss": 0.0949,
2061
+ "step": 1465
2062
+ },
2063
+ {
2064
+ "epoch": 2.1523809523809523,
2065
+ "grad_norm": 0.8748860359191895,
2066
+ "learning_rate": 6.108941901564438e-06,
2067
+ "loss": 0.0965,
2068
+ "step": 1470
2069
+ },
2070
+ {
2071
+ "epoch": 2.1597069597069596,
2072
+ "grad_norm": 0.7672569155693054,
2073
+ "learning_rate": 6.0117161957647945e-06,
2074
+ "loss": 0.0843,
2075
+ "step": 1475
2076
+ },
2077
+ {
2078
+ "epoch": 2.167032967032967,
2079
+ "grad_norm": 1.00459885597229,
2080
+ "learning_rate": 5.915076126956599e-06,
2081
+ "loss": 0.1039,
2082
+ "step": 1480
2083
+ },
2084
+ {
2085
+ "epoch": 2.174358974358974,
2086
+ "grad_norm": 0.7346606254577637,
2087
+ "learning_rate": 5.81902799178116e-06,
2088
+ "loss": 0.0893,
2089
+ "step": 1485
2090
+ },
2091
+ {
2092
+ "epoch": 2.1816849816849815,
2093
+ "grad_norm": 0.832739531993866,
2094
+ "learning_rate": 5.723578048312008e-06,
2095
+ "loss": 0.0963,
2096
+ "step": 1490
2097
+ },
2098
+ {
2099
+ "epoch": 2.1890109890109892,
2100
+ "grad_norm": 0.9014734625816345,
2101
+ "learning_rate": 5.628732515647126e-06,
2102
+ "loss": 0.0858,
2103
+ "step": 1495
2104
+ },
2105
+ {
2106
+ "epoch": 2.1963369963369965,
2107
+ "grad_norm": 0.8096256852149963,
2108
+ "learning_rate": 5.534497573503769e-06,
2109
+ "loss": 0.0945,
2110
+ "step": 1500
2111
+ },
2112
+ {
2113
+ "epoch": 2.203663003663004,
2114
+ "grad_norm": 0.8874215483665466,
2115
+ "learning_rate": 5.4408793618157764e-06,
2116
+ "loss": 0.0962,
2117
+ "step": 1505
2118
+ },
2119
+ {
2120
+ "epoch": 2.210989010989011,
2121
+ "grad_norm": 0.7427526712417603,
2122
+ "learning_rate": 5.34788398033356e-06,
2123
+ "loss": 0.0977,
2124
+ "step": 1510
2125
+ },
2126
+ {
2127
+ "epoch": 2.2183150183150184,
2128
+ "grad_norm": 1.0180338621139526,
2129
+ "learning_rate": 5.255517488226673e-06,
2130
+ "loss": 0.0934,
2131
+ "step": 1515
2132
+ },
2133
+ {
2134
+ "epoch": 2.2256410256410257,
2135
+ "grad_norm": 0.8830599784851074,
2136
+ "learning_rate": 5.163785903688976e-06,
2137
+ "loss": 0.1013,
2138
+ "step": 1520
2139
+ },
2140
+ {
2141
+ "epoch": 2.232967032967033,
2142
+ "grad_norm": 0.8164097666740417,
2143
+ "learning_rate": 5.072695203546564e-06,
2144
+ "loss": 0.0791,
2145
+ "step": 1525
2146
+ },
2147
+ {
2148
+ "epoch": 2.2402930402930403,
2149
+ "grad_norm": 0.9889181852340698,
2150
+ "learning_rate": 4.9822513228683375e-06,
2151
+ "loss": 0.0777,
2152
+ "step": 1530
2153
+ },
2154
+ {
2155
+ "epoch": 2.2476190476190476,
2156
+ "grad_norm": 0.9313451051712036,
2157
+ "learning_rate": 4.892460154579273e-06,
2158
+ "loss": 0.0786,
2159
+ "step": 1535
2160
+ },
2161
+ {
2162
+ "epoch": 2.254945054945055,
2163
+ "grad_norm": 0.7552982568740845,
2164
+ "learning_rate": 4.803327549076488e-06,
2165
+ "loss": 0.0863,
2166
+ "step": 1540
2167
+ },
2168
+ {
2169
+ "epoch": 2.2622710622710622,
2170
+ "grad_norm": 0.8065889477729797,
2171
+ "learning_rate": 4.714859313848047e-06,
2172
+ "loss": 0.0882,
2173
+ "step": 1545
2174
+ },
2175
+ {
2176
+ "epoch": 2.2695970695970695,
2177
+ "grad_norm": 0.8651049733161926,
2178
+ "learning_rate": 4.627061213094576e-06,
2179
+ "loss": 0.0968,
2180
+ "step": 1550
2181
+ },
2182
+ {
2183
+ "epoch": 2.276923076923077,
2184
+ "grad_norm": 0.8217012882232666,
2185
+ "learning_rate": 4.5399389673536896e-06,
2186
+ "loss": 0.0842,
2187
+ "step": 1555
2188
+ },
2189
+ {
2190
+ "epoch": 2.284249084249084,
2191
+ "grad_norm": 0.7754650115966797,
2192
+ "learning_rate": 4.453498253127261e-06,
2193
+ "loss": 0.0866,
2194
+ "step": 1560
2195
+ },
2196
+ {
2197
+ "epoch": 2.2915750915750914,
2198
+ "grad_norm": 0.7794268727302551,
2199
+ "learning_rate": 4.367744702511573e-06,
2200
+ "loss": 0.0841,
2201
+ "step": 1565
2202
+ },
2203
+ {
2204
+ "epoch": 2.2989010989010987,
2205
+ "grad_norm": 0.6935532689094543,
2206
+ "learning_rate": 4.282683902830366e-06,
2207
+ "loss": 0.0835,
2208
+ "step": 1570
2209
+ },
2210
+ {
2211
+ "epoch": 2.306227106227106,
2212
+ "grad_norm": 0.6506497263908386,
2213
+ "learning_rate": 4.198321396270759e-06,
2214
+ "loss": 0.0972,
2215
+ "step": 1575
2216
+ },
2217
+ {
2218
+ "epoch": 2.3135531135531133,
2219
+ "grad_norm": 0.6485122442245483,
2220
+ "learning_rate": 4.114662679522175e-06,
2221
+ "loss": 0.0754,
2222
+ "step": 1580
2223
+ },
2224
+ {
2225
+ "epoch": 2.320879120879121,
2226
+ "grad_norm": 0.7940455079078674,
2227
+ "learning_rate": 4.031713203418208e-06,
2228
+ "loss": 0.0848,
2229
+ "step": 1585
2230
+ },
2231
+ {
2232
+ "epoch": 2.3282051282051284,
2233
+ "grad_norm": 0.8444855213165283,
2234
+ "learning_rate": 3.94947837258142e-06,
2235
+ "loss": 0.0891,
2236
+ "step": 1590
2237
+ },
2238
+ {
2239
+ "epoch": 2.3355311355311357,
2240
+ "grad_norm": 0.7280449271202087,
2241
+ "learning_rate": 3.867963545071273e-06,
2242
+ "loss": 0.0756,
2243
+ "step": 1595
2244
+ },
2245
+ {
2246
+ "epoch": 2.342857142857143,
2247
+ "grad_norm": 0.759167492389679,
2248
+ "learning_rate": 3.7871740320349572e-06,
2249
+ "loss": 0.0889,
2250
+ "step": 1600
2251
+ },
2252
+ {
2253
+ "epoch": 2.3501831501831503,
2254
+ "grad_norm": 0.8187800645828247,
2255
+ "learning_rate": 3.707115097361375e-06,
2256
+ "loss": 0.0881,
2257
+ "step": 1605
2258
+ },
2259
+ {
2260
+ "epoch": 2.3575091575091576,
2261
+ "grad_norm": 0.874309778213501,
2262
+ "learning_rate": 3.6277919573381556e-06,
2263
+ "loss": 0.0856,
2264
+ "step": 1610
2265
+ },
2266
+ {
2267
+ "epoch": 2.364835164835165,
2268
+ "grad_norm": 0.9124408960342407,
2269
+ "learning_rate": 3.5492097803117955e-06,
2270
+ "loss": 0.0784,
2271
+ "step": 1615
2272
+ },
2273
+ {
2274
+ "epoch": 2.372161172161172,
2275
+ "grad_norm": 0.7013418674468994,
2276
+ "learning_rate": 3.4713736863509003e-06,
2277
+ "loss": 0.0834,
2278
+ "step": 1620
2279
+ },
2280
+ {
2281
+ "epoch": 2.3794871794871795,
2282
+ "grad_norm": 0.730495035648346,
2283
+ "learning_rate": 3.3942887469125894e-06,
2284
+ "loss": 0.0815,
2285
+ "step": 1625
2286
+ },
2287
+ {
2288
+ "epoch": 2.3868131868131868,
2289
+ "grad_norm": 0.8347989916801453,
2290
+ "learning_rate": 3.317959984512061e-06,
2291
+ "loss": 0.0811,
2292
+ "step": 1630
2293
+ },
2294
+ {
2295
+ "epoch": 2.394139194139194,
2296
+ "grad_norm": 0.5708280801773071,
2297
+ "learning_rate": 3.2423923723953458e-06,
2298
+ "loss": 0.0666,
2299
+ "step": 1635
2300
+ },
2301
+ {
2302
+ "epoch": 2.4014652014652014,
2303
+ "grad_norm": 0.6894702315330505,
2304
+ "learning_rate": 3.1675908342152852e-06,
2305
+ "loss": 0.0836,
2306
+ "step": 1640
2307
+ },
2308
+ {
2309
+ "epoch": 2.4087912087912087,
2310
+ "grad_norm": 0.9271525740623474,
2311
+ "learning_rate": 3.0935602437106943e-06,
2312
+ "loss": 0.0915,
2313
+ "step": 1645
2314
+ },
2315
+ {
2316
+ "epoch": 2.416117216117216,
2317
+ "grad_norm": 0.6551916003227234,
2318
+ "learning_rate": 3.0203054243888575e-06,
2319
+ "loss": 0.0719,
2320
+ "step": 1650
2321
+ },
2322
+ {
2323
+ "epoch": 2.4234432234432233,
2324
+ "grad_norm": 0.9890082478523254,
2325
+ "learning_rate": 2.947831149211211e-06,
2326
+ "loss": 0.0808,
2327
+ "step": 1655
2328
+ },
2329
+ {
2330
+ "epoch": 2.430769230769231,
2331
+ "grad_norm": 0.660652220249176,
2332
+ "learning_rate": 2.8761421402823638e-06,
2333
+ "loss": 0.0745,
2334
+ "step": 1660
2335
+ },
2336
+ {
2337
+ "epoch": 2.4380952380952383,
2338
+ "grad_norm": 0.7452325820922852,
2339
+ "learning_rate": 2.805243068542456e-06,
2340
+ "loss": 0.0911,
2341
+ "step": 1665
2342
+ },
2343
+ {
2344
+ "epoch": 2.4454212454212456,
2345
+ "grad_norm": 0.6016135215759277,
2346
+ "learning_rate": 2.7351385534627848e-06,
2347
+ "loss": 0.0758,
2348
+ "step": 1670
2349
+ },
2350
+ {
2351
+ "epoch": 2.452747252747253,
2352
+ "grad_norm": 0.7588622570037842,
2353
+ "learning_rate": 2.665833162744837e-06,
2354
+ "loss": 0.0654,
2355
+ "step": 1675
2356
+ },
2357
+ {
2358
+ "epoch": 2.46007326007326,
2359
+ "grad_norm": 0.7467012405395508,
2360
+ "learning_rate": 2.597331412022674e-06,
2361
+ "loss": 0.0775,
2362
+ "step": 1680
2363
+ },
2364
+ {
2365
+ "epoch": 2.4673992673992675,
2366
+ "grad_norm": 0.6321209669113159,
2367
+ "learning_rate": 2.529637764568714e-06,
2368
+ "loss": 0.0781,
2369
+ "step": 1685
2370
+ },
2371
+ {
2372
+ "epoch": 2.474725274725275,
2373
+ "grad_norm": 0.8203696608543396,
2374
+ "learning_rate": 2.46275663100293e-06,
2375
+ "loss": 0.0759,
2376
+ "step": 1690
2377
+ },
2378
+ {
2379
+ "epoch": 2.482051282051282,
2380
+ "grad_norm": 0.6785067915916443,
2381
+ "learning_rate": 2.3966923690054625e-06,
2382
+ "loss": 0.0689,
2383
+ "step": 1695
2384
+ },
2385
+ {
2386
+ "epoch": 2.4893772893772894,
2387
+ "grad_norm": 0.8344903588294983,
2388
+ "learning_rate": 2.3314492830326938e-06,
2389
+ "loss": 0.0885,
2390
+ "step": 1700
2391
+ },
2392
+ {
2393
+ "epoch": 2.4967032967032967,
2394
+ "grad_norm": 0.6424607634544373,
2395
+ "learning_rate": 2.2670316240368095e-06,
2396
+ "loss": 0.0758,
2397
+ "step": 1705
2398
+ },
2399
+ {
2400
+ "epoch": 2.504029304029304,
2401
+ "grad_norm": 0.6638714671134949,
2402
+ "learning_rate": 2.2034435891887935e-06,
2403
+ "loss": 0.0735,
2404
+ "step": 1710
2405
+ },
2406
+ {
2407
+ "epoch": 2.5113553113553113,
2408
+ "grad_norm": 0.7053160071372986,
2409
+ "learning_rate": 2.1406893216049672e-06,
2410
+ "loss": 0.0622,
2411
+ "step": 1715
2412
+ },
2413
+ {
2414
+ "epoch": 2.5186813186813186,
2415
+ "grad_norm": 0.6832835078239441,
2416
+ "learning_rate": 2.0787729100770695e-06,
2417
+ "loss": 0.0793,
2418
+ "step": 1720
2419
+ },
2420
+ {
2421
+ "epoch": 2.526007326007326,
2422
+ "grad_norm": 0.8006079196929932,
2423
+ "learning_rate": 2.0176983888058157e-06,
2424
+ "loss": 0.0761,
2425
+ "step": 1725
2426
+ },
2427
+ {
2428
+ "epoch": 2.533333333333333,
2429
+ "grad_norm": 0.5883076786994934,
2430
+ "learning_rate": 1.9574697371380597e-06,
2431
+ "loss": 0.0686,
2432
+ "step": 1730
2433
+ },
2434
+ {
2435
+ "epoch": 2.5406593406593405,
2436
+ "grad_norm": 0.7064788341522217,
2437
+ "learning_rate": 1.898090879307517e-06,
2438
+ "loss": 0.0764,
2439
+ "step": 1735
2440
+ },
2441
+ {
2442
+ "epoch": 2.547985347985348,
2443
+ "grad_norm": 0.6957082152366638,
2444
+ "learning_rate": 1.839565684179083e-06,
2445
+ "loss": 0.0795,
2446
+ "step": 1740
2447
+ },
2448
+ {
2449
+ "epoch": 2.555311355311355,
2450
+ "grad_norm": 0.7398303747177124,
2451
+ "learning_rate": 1.7818979649967455e-06,
2452
+ "loss": 0.0722,
2453
+ "step": 1745
2454
+ },
2455
+ {
2456
+ "epoch": 2.5626373626373624,
2457
+ "grad_norm": 0.6072902679443359,
2458
+ "learning_rate": 1.7250914791351402e-06,
2459
+ "loss": 0.0688,
2460
+ "step": 1750
2461
+ },
2462
+ {
2463
+ "epoch": 2.5699633699633697,
2464
+ "grad_norm": 0.704990029335022,
2465
+ "learning_rate": 1.6691499278547274e-06,
2466
+ "loss": 0.0838,
2467
+ "step": 1755
2468
+ },
2469
+ {
2470
+ "epoch": 2.577289377289377,
2471
+ "grad_norm": 0.8373236656188965,
2472
+ "learning_rate": 1.6140769560606412e-06,
2473
+ "loss": 0.0751,
2474
+ "step": 1760
2475
+ },
2476
+ {
2477
+ "epoch": 2.5846153846153848,
2478
+ "grad_norm": 0.6890259981155396,
2479
+ "learning_rate": 1.5598761520652105e-06,
2480
+ "loss": 0.0768,
2481
+ "step": 1765
2482
+ },
2483
+ {
2484
+ "epoch": 2.591941391941392,
2485
+ "grad_norm": 0.6583132147789001,
2486
+ "learning_rate": 1.5065510473541278e-06,
2487
+ "loss": 0.0802,
2488
+ "step": 1770
2489
+ },
2490
+ {
2491
+ "epoch": 2.5992673992673994,
2492
+ "grad_norm": 0.6919366121292114,
2493
+ "learning_rate": 1.454105116356404e-06,
2494
+ "loss": 0.0764,
2495
+ "step": 1775
2496
+ },
2497
+ {
2498
+ "epoch": 2.6065934065934067,
2499
+ "grad_norm": 0.6675224304199219,
2500
+ "learning_rate": 1.4025417762179454e-06,
2501
+ "loss": 0.0697,
2502
+ "step": 1780
2503
+ },
2504
+ {
2505
+ "epoch": 2.613919413919414,
2506
+ "grad_norm": 0.6373467445373535,
2507
+ "learning_rate": 1.3518643865789232e-06,
2508
+ "loss": 0.0808,
2509
+ "step": 1785
2510
+ },
2511
+ {
2512
+ "epoch": 2.6212454212454213,
2513
+ "grad_norm": 0.6020922064781189,
2514
+ "learning_rate": 1.3020762493548883e-06,
2515
+ "loss": 0.0635,
2516
+ "step": 1790
2517
+ },
2518
+ {
2519
+ "epoch": 2.6285714285714286,
2520
+ "grad_norm": 0.7419832348823547,
2521
+ "learning_rate": 1.2531806085216085e-06,
2522
+ "loss": 0.082,
2523
+ "step": 1795
2524
+ },
2525
+ {
2526
+ "epoch": 2.635897435897436,
2527
+ "grad_norm": 0.6512110829353333,
2528
+ "learning_rate": 1.2051806499037233e-06,
2529
+ "loss": 0.0761,
2530
+ "step": 1800
2531
+ },
2532
+ {
2533
+ "epoch": 2.643223443223443,
2534
+ "grad_norm": 0.6165404915809631,
2535
+ "learning_rate": 1.1580795009671602e-06,
2536
+ "loss": 0.074,
2537
+ "step": 1805
2538
+ },
2539
+ {
2540
+ "epoch": 2.6505494505494505,
2541
+ "grad_norm": 0.7380002737045288,
2542
+ "learning_rate": 1.1118802306153675e-06,
2543
+ "loss": 0.0872,
2544
+ "step": 1810
2545
+ },
2546
+ {
2547
+ "epoch": 2.6578754578754578,
2548
+ "grad_norm": 0.7050730586051941,
2549
+ "learning_rate": 1.0665858489893565e-06,
2550
+ "loss": 0.0792,
2551
+ "step": 1815
2552
+ },
2553
+ {
2554
+ "epoch": 2.665201465201465,
2555
+ "grad_norm": 0.611894965171814,
2556
+ "learning_rate": 1.0221993072715724e-06,
2557
+ "loss": 0.0812,
2558
+ "step": 1820
2559
+ },
2560
+ {
2561
+ "epoch": 2.672527472527473,
2562
+ "grad_norm": 0.6332575678825378,
2563
+ "learning_rate": 9.787234974936093e-07,
2564
+ "loss": 0.0764,
2565
+ "step": 1825
2566
+ },
2567
+ {
2568
+ "epoch": 2.67985347985348,
2569
+ "grad_norm": 0.6269590854644775,
2570
+ "learning_rate": 9.361612523477847e-07,
2571
+ "loss": 0.0766,
2572
+ "step": 1830
2573
+ },
2574
+ {
2575
+ "epoch": 2.6871794871794874,
2576
+ "grad_norm": 0.7381962537765503,
2577
+ "learning_rate": 8.945153450025617e-07,
2578
+ "loss": 0.0766,
2579
+ "step": 1835
2580
+ },
2581
+ {
2582
+ "epoch": 2.6945054945054947,
2583
+ "grad_norm": 0.6251266002655029,
2584
+ "learning_rate": 8.537884889218639e-07,
2585
+ "loss": 0.0717,
2586
+ "step": 1840
2587
+ },
2588
+ {
2589
+ "epoch": 2.701831501831502,
2590
+ "grad_norm": 0.61935955286026,
2591
+ "learning_rate": 8.139833376882932e-07,
2592
+ "loss": 0.0756,
2593
+ "step": 1845
2594
+ },
2595
+ {
2596
+ "epoch": 2.7091575091575093,
2597
+ "grad_norm": 0.6990112662315369,
2598
+ "learning_rate": 7.751024848302135e-07,
2599
+ "loss": 0.0758,
2600
+ "step": 1850
2601
+ },
2602
+ {
2603
+ "epoch": 2.7164835164835166,
2604
+ "grad_norm": 0.661233127117157,
2605
+ "learning_rate": 7.371484636527803e-07,
2606
+ "loss": 0.0698,
2607
+ "step": 1855
2608
+ },
2609
+ {
2610
+ "epoch": 2.723809523809524,
2611
+ "grad_norm": 0.6670806407928467,
2612
+ "learning_rate": 7.001237470728749e-07,
2613
+ "loss": 0.0828,
2614
+ "step": 1860
2615
+ },
2616
+ {
2617
+ "epoch": 2.731135531135531,
2618
+ "grad_norm": 0.5686982274055481,
2619
+ "learning_rate": 6.640307474579898e-07,
2620
+ "loss": 0.068,
2621
+ "step": 1865
2622
+ },
2623
+ {
2624
+ "epoch": 2.7384615384615385,
2625
+ "grad_norm": 0.7640694975852966,
2626
+ "learning_rate": 6.288718164690344e-07,
2627
+ "loss": 0.08,
2628
+ "step": 1870
2629
+ },
2630
+ {
2631
+ "epoch": 2.745787545787546,
2632
+ "grad_norm": 0.6887659430503845,
2633
+ "learning_rate": 5.946492449071289e-07,
2634
+ "loss": 0.0628,
2635
+ "step": 1875
2636
+ },
2637
+ {
2638
+ "epoch": 2.753113553113553,
2639
+ "grad_norm": 0.6388258934020996,
2640
+ "learning_rate": 5.613652625643279e-07,
2641
+ "loss": 0.0735,
2642
+ "step": 1880
2643
+ },
2644
+ {
2645
+ "epoch": 2.7604395604395604,
2646
+ "grad_norm": 0.6812204718589783,
2647
+ "learning_rate": 5.290220380783539e-07,
2648
+ "loss": 0.0717,
2649
+ "step": 1885
2650
+ },
2651
+ {
2652
+ "epoch": 2.7677655677655677,
2653
+ "grad_norm": 0.6990679502487183,
2654
+ "learning_rate": 4.976216787912846e-07,
2655
+ "loss": 0.0741,
2656
+ "step": 1890
2657
+ },
2658
+ {
2659
+ "epoch": 2.775091575091575,
2660
+ "grad_norm": 0.7488406896591187,
2661
+ "learning_rate": 4.6716623061225204e-07,
2662
+ "loss": 0.0706,
2663
+ "step": 1895
2664
+ },
2665
+ {
2666
+ "epoch": 2.7824175824175823,
2667
+ "grad_norm": 0.6586956977844238,
2668
+ "learning_rate": 4.3765767788414957e-07,
2669
+ "loss": 0.0691,
2670
+ "step": 1900
2671
+ },
2672
+ {
2673
+ "epoch": 2.7897435897435896,
2674
+ "grad_norm": 0.6575021147727966,
2675
+ "learning_rate": 4.0909794325432517e-07,
2676
+ "loss": 0.0638,
2677
+ "step": 1905
2678
+ },
2679
+ {
2680
+ "epoch": 2.797069597069597,
2681
+ "grad_norm": 0.7643905878067017,
2682
+ "learning_rate": 3.8148888754931997e-07,
2683
+ "loss": 0.07,
2684
+ "step": 1910
2685
+ },
2686
+ {
2687
+ "epoch": 2.804395604395604,
2688
+ "grad_norm": 0.7247445583343506,
2689
+ "learning_rate": 3.548323096536238e-07,
2690
+ "loss": 0.0817,
2691
+ "step": 1915
2692
+ },
2693
+ {
2694
+ "epoch": 2.8117216117216115,
2695
+ "grad_norm": 0.5048761367797852,
2696
+ "learning_rate": 3.291299463924652e-07,
2697
+ "loss": 0.068,
2698
+ "step": 1920
2699
+ },
2700
+ {
2701
+ "epoch": 2.819047619047619,
2702
+ "grad_norm": 0.6676440834999084,
2703
+ "learning_rate": 3.0438347241865225e-07,
2704
+ "loss": 0.066,
2705
+ "step": 1925
2706
+ },
2707
+ {
2708
+ "epoch": 2.826373626373626,
2709
+ "grad_norm": 0.6631476283073425,
2710
+ "learning_rate": 2.8059450010345323e-07,
2711
+ "loss": 0.0726,
2712
+ "step": 1930
2713
+ },
2714
+ {
2715
+ "epoch": 2.833699633699634,
2716
+ "grad_norm": 0.5443286895751953,
2717
+ "learning_rate": 2.57764579431542e-07,
2718
+ "loss": 0.0715,
2719
+ "step": 1935
2720
+ },
2721
+ {
2722
+ "epoch": 2.841025641025641,
2723
+ "grad_norm": 0.4935495853424072,
2724
+ "learning_rate": 2.3589519790002233e-07,
2725
+ "loss": 0.0675,
2726
+ "step": 1940
2727
+ },
2728
+ {
2729
+ "epoch": 2.8483516483516484,
2730
+ "grad_norm": 0.5985262393951416,
2731
+ "learning_rate": 2.149877804214856e-07,
2732
+ "loss": 0.0732,
2733
+ "step": 1945
2734
+ },
2735
+ {
2736
+ "epoch": 2.8556776556776557,
2737
+ "grad_norm": 0.6776986122131348,
2738
+ "learning_rate": 1.95043689231183e-07,
2739
+ "loss": 0.0771,
2740
+ "step": 1950
2741
+ },
2742
+ {
2743
+ "epoch": 2.863003663003663,
2744
+ "grad_norm": 0.6605568528175354,
2745
+ "learning_rate": 1.7606422379827513e-07,
2746
+ "loss": 0.0621,
2747
+ "step": 1955
2748
+ },
2749
+ {
2750
+ "epoch": 2.8703296703296703,
2751
+ "grad_norm": 0.5364888906478882,
2752
+ "learning_rate": 1.5805062074114807e-07,
2753
+ "loss": 0.0655,
2754
+ "step": 1960
2755
+ },
2756
+ {
2757
+ "epoch": 2.8776556776556776,
2758
+ "grad_norm": 0.6770036220550537,
2759
+ "learning_rate": 1.4100405374684932e-07,
2760
+ "loss": 0.0781,
2761
+ "step": 1965
2762
+ },
2763
+ {
2764
+ "epoch": 2.884981684981685,
2765
+ "grad_norm": 0.560341477394104,
2766
+ "learning_rate": 1.249256334946225e-07,
2767
+ "loss": 0.0724,
2768
+ "step": 1970
2769
+ },
2770
+ {
2771
+ "epoch": 2.8923076923076922,
2772
+ "grad_norm": 0.7277321815490723,
2773
+ "learning_rate": 1.0981640758352174e-07,
2774
+ "loss": 0.0804,
2775
+ "step": 1975
2776
+ },
2777
+ {
2778
+ "epoch": 2.8996336996336995,
2779
+ "grad_norm": 0.6494090557098389,
2780
+ "learning_rate": 9.567736046417641e-08,
2781
+ "loss": 0.0701,
2782
+ "step": 1980
2783
+ },
2784
+ {
2785
+ "epoch": 2.906959706959707,
2786
+ "grad_norm": 0.5981951951980591,
2787
+ "learning_rate": 8.25094133746307e-08,
2788
+ "loss": 0.071,
2789
+ "step": 1985
2790
+ },
2791
+ {
2792
+ "epoch": 2.914285714285714,
2793
+ "grad_norm": 0.5359269380569458,
2794
+ "learning_rate": 7.031342428032994e-08,
2795
+ "loss": 0.064,
2796
+ "step": 1990
2797
+ },
2798
+ {
2799
+ "epoch": 2.9216117216117214,
2800
+ "grad_norm": 0.6491533517837524,
2801
+ "learning_rate": 5.909018781821873e-08,
2802
+ "loss": 0.0778,
2803
+ "step": 1995
2804
+ },
2805
+ {
2806
+ "epoch": 2.928937728937729,
2807
+ "grad_norm": 0.584261417388916,
2808
+ "learning_rate": 4.884043524496063e-08,
2809
+ "loss": 0.0711,
2810
+ "step": 2000
2811
+ },
2812
+ {
2813
+ "epoch": 2.9362637362637365,
2814
+ "grad_norm": 0.6126725077629089,
2815
+ "learning_rate": 3.956483438929459e-08,
2816
+ "loss": 0.0698,
2817
+ "step": 2005
2818
+ },
2819
+ {
2820
+ "epoch": 2.943589743589744,
2821
+ "grad_norm": 0.5821678638458252,
2822
+ "learning_rate": 3.126398960852983e-08,
2823
+ "loss": 0.0777,
2824
+ "step": 2010
2825
+ },
2826
+ {
2827
+ "epoch": 2.950915750915751,
2828
+ "grad_norm": 0.586898922920227,
2829
+ "learning_rate": 2.393844174915727e-08,
2830
+ "loss": 0.0655,
2831
+ "step": 2015
2832
+ },
2833
+ {
2834
+ "epoch": 2.9582417582417584,
2835
+ "grad_norm": 0.6654184460639954,
2836
+ "learning_rate": 1.7588668111614414e-08,
2837
+ "loss": 0.0701,
2838
+ "step": 2020
2839
+ },
2840
+ {
2841
+ "epoch": 2.9655677655677657,
2842
+ "grad_norm": 0.6260083913803101,
2843
+ "learning_rate": 1.2215082419190227e-08,
2844
+ "loss": 0.0719,
2845
+ "step": 2025
2846
+ },
2847
+ {
2848
+ "epoch": 2.972893772893773,
2849
+ "grad_norm": 0.7631379961967468,
2850
+ "learning_rate": 7.81803479106502e-09,
2851
+ "loss": 0.0794,
2852
+ "step": 2030
2853
+ },
2854
+ {
2855
+ "epoch": 2.9802197802197803,
2856
+ "grad_norm": 0.5739051699638367,
2857
+ "learning_rate": 4.397811719500378e-09,
2858
+ "loss": 0.0681,
2859
+ "step": 2035
2860
+ },
2861
+ {
2862
+ "epoch": 2.9875457875457876,
2863
+ "grad_norm": 0.7282893061637878,
2864
+ "learning_rate": 1.9546360511690785e-09,
2865
+ "loss": 0.0863,
2866
+ "step": 2040
2867
+ },
2868
+ {
2869
+ "epoch": 2.994871794871795,
2870
+ "grad_norm": 0.5536500811576843,
2871
+ "learning_rate": 4.886669726400506e-10,
2872
+ "loss": 0.0694,
2873
+ "step": 2045
2874
+ },
2875
+ {
2876
+ "epoch": 3.0,
2877
+ "step": 2049,
2878
+ "total_flos": 2.6763548800791347e+18,
2879
+ "train_loss": 0.38947603013411913,
2880
+ "train_runtime": 1269.3059,
2881
+ "train_samples_per_second": 51.583,
2882
+ "train_steps_per_second": 1.614
2883
+ }
2884
+ ],
2885
+ "logging_steps": 5,
2886
+ "max_steps": 2049,
2887
+ "num_input_tokens_seen": 0,
2888
+ "num_train_epochs": 3,
2889
+ "save_steps": 20000,
2890
+ "stateful_callbacks": {
2891
+ "TrainerControl": {
2892
+ "args": {
2893
+ "should_epoch_stop": false,
2894
+ "should_evaluate": false,
2895
+ "should_log": false,
2896
+ "should_save": false,
2897
+ "should_training_stop": false
2898
+ },
2899
+ "attributes": {}
2900
+ }
2901
+ },
2902
+ "total_flos": 2.6763548800791347e+18,
2903
+ "train_batch_size": 2,
2904
+ "trial_name": null,
2905
+ "trial_params": null
2906
+ }
instruct/134_128_e3_3e-5/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3784d91f391cb0556a373e6a50c07ef546d3483f560c3d0168101b5cdd85372a
3
+ size 8273
instruct/134_128_e3_3e-5/vocab.json ADDED
The diff for this file is too large to render. See raw diff