Hugging Face's logo

melihcatal
/
codedp-cpt-models-v2

PEFT
Safetensors
code-generation
differential-privacy
continued-pretraining
lora
Model card Files
xet
 Community
melihcatal commited on
Commit
1a99fa7
·
verified ·
1 Parent(s): 080080b

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. granite-4.0-h-tiny/base_attn/adapter/README.md +207 -0
  2. granite-4.0-h-tiny/base_attn/adapter/adapter_config.json +46 -0
  3. granite-4.0-h-tiny/base_attn/audit_results.json +137 -0
  4. granite-4.0-h-tiny/base_attn/canary_meta.json +0 -0
  5. granite-4.0-h-tiny/base_attn/metrics.jsonl +204 -0
  6. granite-4.0-h-tiny/base_attn/resolved_config.yaml +111 -0
  7. granite-4.0-h-tiny/base_attn/summary.json +71 -0
  8. granite-4.0-h-tiny/base_attn/tokenizer/tokenizer.json +0 -0
  9. granite-4.0-h-tiny/base_attn/tokenizer/tokenizer_config.json +516 -0
  10. granite-4.0-h-tiny/base_attn/train.log +195 -0
  11. granite-4.0-h-tiny/dp3_attn/adapter/README.md +207 -0
  12. granite-4.0-h-tiny/dp3_attn/adapter/adapter_config.json +46 -0
  13. granite-4.0-h-tiny/dp3_attn/audit_results.json +137 -0
  14. granite-4.0-h-tiny/dp3_attn/canary_meta.json +0 -0
  15. granite-4.0-h-tiny/dp3_attn/metrics.jsonl +0 -0
  16. granite-4.0-h-tiny/dp3_attn/resolved_config.yaml +110 -0
  17. granite-4.0-h-tiny/dp3_attn/summary.json +72 -0
  18. granite-4.0-h-tiny/dp3_attn/tokenizer/tokenizer.json +0 -0
  19. granite-4.0-h-tiny/dp3_attn/tokenizer/tokenizer_config.json +516 -0
  20. granite-4.0-h-tiny/dp3_attn/train.log +196 -0
  21. granite-4.0-h-tiny/dp8_attn/adapter/README.md +207 -0
  22. granite-4.0-h-tiny/dp8_attn/adapter/adapter_config.json +46 -0
  23. granite-4.0-h-tiny/dp8_attn/audit_results.json +137 -0
  24. granite-4.0-h-tiny/dp8_attn/canary_meta.json +0 -0
  25. granite-4.0-h-tiny/dp8_attn/metrics.jsonl +0 -0
  26. granite-4.0-h-tiny/dp8_attn/resolved_config.yaml +111 -0
  27. granite-4.0-h-tiny/dp8_attn/summary.json +72 -0
  28. granite-4.0-h-tiny/dp8_attn/tokenizer/tokenizer.json +0 -0
  29. granite-4.0-h-tiny/dp8_attn/tokenizer/tokenizer_config.json +516 -0
  30. granite-4.0-h-tiny/dp8_attn/train.log +196 -0
  31. llama3.1-8b/dp3/adapter/README.md +207 -0
  32. llama3.1-8b/dp3/adapter/adapter_config.json +46 -0
  33. llama3.1-8b/dp3/audit_results.json +137 -0
  34. llama3.1-8b/dp3/canary_meta.json +0 -0
  35. llama3.1-8b/dp3/metrics.jsonl +0 -0
  36. llama3.1-8b/dp3/resolved_config.yaml +102 -0
  37. llama3.1-8b/dp3/summary.json +72 -0
  38. llama3.1-8b/dp3/tokenizer/tokenizer_config.json +516 -0
  39. llama3.1-8b/dp3/train.log +276 -0
  40. llama3.1-8b/dp8/adapter/README.md +207 -0
  41. llama3.1-8b/dp8/adapter/adapter_config.json +46 -0
  42. llama3.1-8b/dp8/audit_results.json +137 -0
  43. llama3.1-8b/dp8/canary_meta.json +0 -0
  44. llama3.1-8b/dp8/metrics.jsonl +0 -0
  45. llama3.1-8b/dp8/resolved_config.yaml +102 -0
  46. llama3.1-8b/dp8/summary.json +72 -0
  47. llama3.1-8b/dp8/tokenizer/tokenizer_config.json +516 -0
  48. llama3.1-8b/dp8/train.log +276 -0
  49. qwen1.5-moe-a2.7b/dp3_attn_v2/adapter/adapter_config.json +46 -0
  50. qwen1.5-moe-a2.7b/dp3_attn_v2/tokenizer/merges.txt +0 -0
granite-4.0-h-tiny/base_attn/adapter/README.md ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ibm-granite/granite-4.0-h-tiny-base
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:ibm-granite/granite-4.0-h-tiny-base
7
+ - lora
8
+ - transformers
9
+ ---
10
+
11
+ # Model Card for Model ID
12
+
13
+ <!-- Provide a quick summary of what the model is/does. -->
14
+
15
+
16
+
17
+ ## Model Details
18
+
19
+ ### Model Description
20
+
21
+ <!-- Provide a longer summary of what this model is. -->
22
+
23
+
24
+
25
+ - **Developed by:** [More Information Needed]
26
+ - **Funded by [optional]:** [More Information Needed]
27
+ - **Shared by [optional]:** [More Information Needed]
28
+ - **Model type:** [More Information Needed]
29
+ - **Language(s) (NLP):** [More Information Needed]
30
+ - **License:** [More Information Needed]
31
+ - **Finetuned from model [optional]:** [More Information Needed]
32
+
33
+ ### Model Sources [optional]
34
+
35
+ <!-- Provide the basic links for the model. -->
36
+
37
+ - **Repository:** [More Information Needed]
38
+ - **Paper [optional]:** [More Information Needed]
39
+ - **Demo [optional]:** [More Information Needed]
40
+
41
+ ## Uses
42
+
43
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
44
+
45
+ ### Direct Use
46
+
47
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
48
+
49
+ [More Information Needed]
50
+
51
+ ### Downstream Use [optional]
52
+
53
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
54
+
55
+ [More Information Needed]
56
+
57
+ ### Out-of-Scope Use
58
+
59
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
60
+
61
+ [More Information Needed]
62
+
63
+ ## Bias, Risks, and Limitations
64
+
65
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
66
+
67
+ [More Information Needed]
68
+
69
+ ### Recommendations
70
+
71
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
72
+
73
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
74
+
75
+ ## How to Get Started with the Model
76
+
77
+ Use the code below to get started with the model.
78
+
79
+ [More Information Needed]
80
+
81
+ ## Training Details
82
+
83
+ ### Training Data
84
+
85
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
86
+
87
+ [More Information Needed]
88
+
89
+ ### Training Procedure
90
+
91
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
92
+
93
+ #### Preprocessing [optional]
94
+
95
+ [More Information Needed]
96
+
97
+
98
+ #### Training Hyperparameters
99
+
100
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
101
+
102
+ #### Speeds, Sizes, Times [optional]
103
+
104
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
105
+
106
+ [More Information Needed]
107
+
108
+ ## Evaluation
109
+
110
+ <!-- This section describes the evaluation protocols and provides the results. -->
111
+
112
+ ### Testing Data, Factors & Metrics
113
+
114
+ #### Testing Data
115
+
116
+ <!-- This should link to a Dataset Card if possible. -->
117
+
118
+ [More Information Needed]
119
+
120
+ #### Factors
121
+
122
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
123
+
124
+ [More Information Needed]
125
+
126
+ #### Metrics
127
+
128
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
129
+
130
+ [More Information Needed]
131
+
132
+ ### Results
133
+
134
+ [More Information Needed]
135
+
136
+ #### Summary
137
+
138
+
139
+
140
+ ## Model Examination [optional]
141
+
142
+ <!-- Relevant interpretability work for the model goes here -->
143
+
144
+ [More Information Needed]
145
+
146
+ ## Environmental Impact
147
+
148
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
149
+
150
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
151
+
152
+ - **Hardware Type:** [More Information Needed]
153
+ - **Hours used:** [More Information Needed]
154
+ - **Cloud Provider:** [More Information Needed]
155
+ - **Compute Region:** [More Information Needed]
156
+ - **Carbon Emitted:** [More Information Needed]
157
+
158
+ ## Technical Specifications [optional]
159
+
160
+ ### Model Architecture and Objective
161
+
162
+ [More Information Needed]
163
+
164
+ ### Compute Infrastructure
165
+
166
+ [More Information Needed]
167
+
168
+ #### Hardware
169
+
170
+ [More Information Needed]
171
+
172
+ #### Software
173
+
174
+ [More Information Needed]
175
+
176
+ ## Citation [optional]
177
+
178
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
179
+
180
+ **BibTeX:**
181
+
182
+ [More Information Needed]
183
+
184
+ **APA:**
185
+
186
+ [More Information Needed]
187
+
188
+ ## Glossary [optional]
189
+
190
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
191
+
192
+ [More Information Needed]
193
+
194
+ ## More Information [optional]
195
+
196
+ [More Information Needed]
197
+
198
+ ## Model Card Authors [optional]
199
+
200
+ [More Information Needed]
201
+
202
+ ## Model Card Contact
203
+
204
+ [More Information Needed]
205
+ ### Framework versions
206
+
207
+ - PEFT 0.18.1
granite-4.0-h-tiny/base_attn/adapter/adapter_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "ibm-granite/granite-4.0-h-tiny-base",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": true,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": [
25
+ "lm_head",
26
+ "embed_tokens"
27
+ ],
28
+ "peft_type": "LORA",
29
+ "peft_version": "0.18.1",
30
+ "qalora_group_size": 16,
31
+ "r": 16,
32
+ "rank_pattern": {},
33
+ "revision": null,
34
+ "target_modules": [
35
+ "q_proj",
36
+ "o_proj",
37
+ "v_proj",
38
+ "k_proj"
39
+ ],
40
+ "target_parameters": null,
41
+ "task_type": "CAUSAL_LM",
42
+ "trainable_token_indices": null,
43
+ "use_dora": false,
44
+ "use_qalora": false,
45
+ "use_rslora": false
46
+ }
granite-4.0-h-tiny/base_attn/audit_results.json ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "delta": 1e-05,
3
+ "num_canaries": 500,
4
+ "num_members": 250,
5
+ "paper_guess_fraction": 0.2,
6
+ "paper_guess_steps": 20,
7
+ "loss": {
8
+ "auc": 0.922912,
9
+ "empirical_epsilon": {
10
+ "0.05": 3.4791953936219215,
11
+ "0.01": 3.023197554051876
12
+ },
13
+ "empirical_epsilon_details": {
14
+ "0.05": {
15
+ "epsilon": 3.4791953936219215,
16
+ "num_guesses": 100,
17
+ "correct_guesses": 100,
18
+ "candidate_num_guesses": [
19
+ 5,
20
+ 10,
21
+ 15,
22
+ 20,
23
+ 25,
24
+ 30,
25
+ 35,
26
+ 40,
27
+ 45,
28
+ 50,
29
+ 55,
30
+ 60,
31
+ 65,
32
+ 70,
33
+ 75,
34
+ 80,
35
+ 85,
36
+ 90,
37
+ 95,
38
+ 100
39
+ ],
40
+ "direction": "lower"
41
+ },
42
+ "0.01": {
43
+ "epsilon": 3.023197554051876,
44
+ "num_guesses": 100,
45
+ "correct_guesses": 100,
46
+ "candidate_num_guesses": [
47
+ 5,
48
+ 10,
49
+ 15,
50
+ 20,
51
+ 25,
52
+ 30,
53
+ 35,
54
+ 40,
55
+ 45,
56
+ 50,
57
+ 55,
58
+ 60,
59
+ 65,
60
+ 70,
61
+ 75,
62
+ 80,
63
+ 85,
64
+ 90,
65
+ 95,
66
+ 100
67
+ ],
68
+ "direction": "lower"
69
+ }
70
+ }
71
+ },
72
+ "embedding": {
73
+ "auc": 0.756,
74
+ "empirical_epsilon": {
75
+ "0.05": 3.4791953936219215,
76
+ "0.01": 3.023197554051876
77
+ },
78
+ "empirical_epsilon_details": {
79
+ "0.05": {
80
+ "epsilon": 3.4791953936219215,
81
+ "num_guesses": 100,
82
+ "correct_guesses": 100,
83
+ "candidate_num_guesses": [
84
+ 5,
85
+ 10,
86
+ 15,
87
+ 20,
88
+ 25,
89
+ 30,
90
+ 35,
91
+ 40,
92
+ 45,
93
+ 50,
94
+ 55,
95
+ 60,
96
+ 65,
97
+ 70,
98
+ 75,
99
+ 80,
100
+ 85,
101
+ 90,
102
+ 95,
103
+ 100
104
+ ],
105
+ "direction": "lower"
106
+ },
107
+ "0.01": {
108
+ "epsilon": 3.023197554051876,
109
+ "num_guesses": 100,
110
+ "correct_guesses": 100,
111
+ "candidate_num_guesses": [
112
+ 5,
113
+ 10,
114
+ 15,
115
+ 20,
116
+ 25,
117
+ 30,
118
+ 35,
119
+ 40,
120
+ 45,
121
+ 50,
122
+ 55,
123
+ 60,
124
+ 65,
125
+ 70,
126
+ 75,
127
+ 80,
128
+ 85,
129
+ 90,
130
+ 95,
131
+ 100
132
+ ],
133
+ "direction": "lower"
134
+ }
135
+ }
136
+ }
137
+ }
granite-4.0-h-tiny/base_attn/canary_meta.json ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/base_attn/metrics.jsonl ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"timestamp": 1774538234.7641191, "event": "energy_final", "step": 0, "epoch": null, "metrics": {"energy/codecarbon/duration": 159.4880690809805, "energy/codecarbon/emissions": 0.0017978973401347282, "energy/codecarbon/emissions_rate": 1.1272926874685786e-05, "energy/codecarbon/cpu_power": 80.51921966218183, "energy/codecarbon/gpu_power": 1089.4685309350648, "energy/codecarbon/ram_power": 38.0, "energy/codecarbon/cpu_energy": 0.003416534547822658, "energy/codecarbon/gpu_energy": 0.046571110312413566, "energy/codecarbon/ram_energy": 0.0016123132413832675, "energy/codecarbon/energy_consumed": 0.05159995810161949, "energy/codecarbon/water_consumed": 0.0, "energy/codecarbon/cpu_count": 8.0, "energy/codecarbon/gpu_count": 8.0, "energy/codecarbon/longitude": 8.212, "energy/codecarbon/latitude": 47.4843, "energy/codecarbon/ram_total_size": 256.0, "energy/codecarbon/cpu_utilization_percent": 10.445222929936307, "energy/codecarbon/gpu_utilization_percent": 8.926751592356688, "energy/codecarbon/ram_utilization_percent": 25.803821656050953, "energy/codecarbon/ram_used_gb": 511.3191909790039, "energy/codecarbon/pue": 1.0, "energy/codecarbon/wue": 0.0}}
2
+ {"timestamp": 1774538505.1569417, "event": "train_step", "step": 5, "epoch": 1, "metrics": {"train/step_loss": 3.1499931812286377, "train/step_real_loss": 3.1499931812286377, "train/lr": 2.576606060606061e-05, "perf/step_duration_sec": 5.200812005903572, "perf/samples_per_sec": 6.1528853501483995, "perf/tokens_per_sec": 5249.564869679738, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27302.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 89.7648868560791}}
3
+ {"timestamp": 1774538528.2468174, "event": "train_step", "step": 10, "epoch": 1, "metrics": {"train/step_loss": 3.2097513675689697, "train/step_real_loss": 3.2097513675689697, "train/lr": 5.152212121212123e-05, "perf/step_duration_sec": 4.318874393124133, "perf/samples_per_sec": 7.40933796336972, "perf/tokens_per_sec": 6131.6902483111535, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26482.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 89.7648868560791}}
4
+ {"timestamp": 1774538544.3442416, "event": "eval_step", "step": 10, "epoch": 1, "metrics": {"eval/loss": 3.1471739139245907, "eval/duration_sec": 16.087412405991927}}
5
+ {"timestamp": 1774538563.7449312, "event": "train_step", "step": 15, "epoch": 1, "metrics": {"train/step_loss": 2.908831000328064, "train/step_real_loss": 2.908831000328064, "train/lr": 7.727818181818186e-05, "perf/step_duration_sec": 3.7661584180314094, "perf/samples_per_sec": 8.496721711649762, "perf/tokens_per_sec": 7373.826832944549, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27771.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
6
+ {"timestamp": 1774538582.3235736, "event": "train_step", "step": 20, "epoch": 1, "metrics": {"train/step_loss": 2.7785194064631606, "train/step_real_loss": 2.3614416122436523, "train/lr": 0.00010303424242424251, "train/step_canary_loss": 16.125, "perf/step_duration_sec": 3.6936976718716323, "perf/samples_per_sec": 8.934136719229265, "perf/tokens_per_sec": 7095.87040639391, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 26210.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
7
+ {"timestamp": 1774538599.1987338, "event": "eval_step", "step": 20, "epoch": 1, "metrics": {"eval/loss": 2.0248701695514764, "eval/duration_sec": 16.85699842683971}}
8
+ {"timestamp": 1774538617.333866, "event": "train_step", "step": 25, "epoch": 1, "metrics": {"train/step_loss": 1.6351239681243896, "train/step_real_loss": 1.6351239681243896, "train/lr": 0.00012879030303030313, "perf/step_duration_sec": 3.4865305959247053, "perf/samples_per_sec": 9.17817845551213, "perf/tokens_per_sec": 8053.56477663516, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28079.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
9
+ {"timestamp": 1774538635.413711, "event": "train_step", "step": 30, "epoch": 1, "metrics": {"train/step_loss": 1.2344295382499695, "train/step_real_loss": 1.2344295382499695, "train/lr": 0.00015454636363636373, "perf/step_duration_sec": 3.5152053399942815, "perf/samples_per_sec": 9.103308883813899, "perf/tokens_per_sec": 7756.303647412061, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27265.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
10
+ {"timestamp": 1774538652.0878432, "event": "eval_step", "step": 30, "epoch": 1, "metrics": {"eval/loss": 1.1229395265488522, "eval/duration_sec": 16.65939668309875}}
11
+ {"timestamp": 1774538670.993655, "event": "train_step", "step": 35, "epoch": 1, "metrics": {"train/step_loss": 1.027151733636856, "train/step_real_loss": 1.027151733636856, "train/lr": 0.00016999586681003042, "perf/step_duration_sec": 3.514393205055967, "perf/samples_per_sec": 9.10541255143657, "perf/tokens_per_sec": 7394.16408005096, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25986.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
12
+ {"timestamp": 1774538688.9048922, "event": "train_step", "step": 40, "epoch": 1, "metrics": {"train/step_loss": 1.0080198645591736, "train/step_real_loss": 1.0080198645591736, "train/lr": 0.0001699493732819447, "perf/step_duration_sec": 3.527619219152257, "perf/samples_per_sec": 9.071273856958436, "perf/tokens_per_sec": 7355.952666069202, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25949.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
13
+ {"timestamp": 1774538705.3989658, "event": "eval_step", "step": 40, "epoch": 1, "metrics": {"eval/loss": 0.916038467184357, "eval/duration_sec": 16.485946608008817}}
14
+ {"timestamp": 1774538722.9317694, "event": "train_step", "step": 45, "epoch": 1, "metrics": {"train/step_loss": 1.3513351823344375, "train/step_real_loss": 0.942392498254776, "train/lr": 0.00016985124958329646, "train/step_canary_loss": 14.4375, "perf/step_duration_sec": 3.5422571569215506, "perf/samples_per_sec": 9.316093817615185, "perf/tokens_per_sec": 8276.64359226794, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 29318.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.583967685699463, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
15
+ {"timestamp": 1774538740.2548163, "event": "train_step", "step": 50, "epoch": 1, "metrics": {"train/step_loss": 1.3081048943779685, "train/step_real_loss": 0.8646081686019897, "train/lr": 0.00016970155849206137, "train/step_canary_loss": 15.5, "perf/step_duration_sec": 3.4046929921023548, "perf/samples_per_sec": 9.692503869379106, "perf/tokens_per_sec": 7358.078998051072, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 25052.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
16
+ {"timestamp": 1774538756.7995102, "event": "eval_step", "step": 50, "epoch": 1, "metrics": {"eval/loss": 0.8713312087499577, "eval/duration_sec": 16.533886903896928}}
17
+ {"timestamp": 1774538774.2574244, "event": "train_step", "step": 55, "epoch": 1, "metrics": {"train/step_loss": 1.3067277308666345, "train/step_real_loss": 0.9413129687309265, "train/lr": 0.0001695003957782096, "train/step_canary_loss": 13.0, "perf/step_duration_sec": 3.5630873620975763, "perf/samples_per_sec": 9.26163089657533, "perf/tokens_per_sec": 7761.52734681184, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 27655.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.583967685699463, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
18
+ {"timestamp": 1774538791.4851599, "event": "train_step", "step": 60, "epoch": 1, "metrics": {"train/step_loss": 0.8761611878871918, "train/step_real_loss": 0.8761611878871918, "train/lr": 0.00016924789014243367, "perf/step_duration_sec": 3.4246954480186105, "perf/samples_per_sec": 9.343896555389735, "perf/tokens_per_sec": 7645.351359680295, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26183.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
19
+ {"timestamp": 1774538808.448436, "event": "eval_step", "step": 60, "epoch": 1, "metrics": {"eval/loss": 0.847663000549959, "eval/duration_sec": 16.94123208615929}}
20
+ {"timestamp": 1774538826.065222, "event": "train_step", "step": 65, "epoch": 1, "metrics": {"train/step_loss": 0.8556468188762665, "train/step_real_loss": 0.8556468188762665, "train/lr": 0.0001689442031338078, "perf/step_duration_sec": 3.377781799994409, "perf/samples_per_sec": 9.473672929391995, "perf/tokens_per_sec": 8310.18747275104, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28070.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
21
+ {"timestamp": 1774538843.1884878, "event": "train_step", "step": 70, "epoch": 1, "metrics": {"train/step_loss": 0.8897512555122375, "train/step_real_loss": 0.8897512555122375, "train/lr": 0.0001685895290464314, "perf/step_duration_sec": 3.293713756138459, "perf/samples_per_sec": 9.715476926421413, "perf/tokens_per_sec": 7986.729250826303, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26306.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
22
+ {"timestamp": 1774538859.5539024, "event": "eval_step", "step": 70, "epoch": 1, "metrics": {"eval/loss": 0.8339212757089863, "eval/duration_sec": 16.358918135985732}}
23
+ {"timestamp": 1774538877.081425, "event": "train_step", "step": 75, "epoch": 1, "metrics": {"train/step_loss": 0.8878340125083923, "train/step_real_loss": 0.8878340125083923, "train/lr": 0.0001681840947951227, "perf/step_duration_sec": 3.504006218863651, "perf/samples_per_sec": 9.13240388322644, "perf/tokens_per_sec": 7742.280779626564, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27129.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
24
+ {"timestamp": 1774538894.4030476, "event": "train_step", "step": 80, "epoch": 1, "metrics": {"train/step_loss": 0.831062525510788, "train/step_real_loss": 0.831062525510788, "train/lr": 0.00016772815977024252, "perf/step_duration_sec": 3.464300711173564, "perf/samples_per_sec": 9.237073414784394, "perf/tokens_per_sec": 8140.45960532296, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28201.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
25
+ {"timestamp": 1774538911.1440206, "event": "eval_step", "step": 80, "epoch": 1, "metrics": {"eval/loss": 0.825627386893915, "eval/duration_sec": 16.71691776998341}}
26
+ {"timestamp": 1774538928.716251, "event": "train_step", "step": 85, "epoch": 1, "metrics": {"train/step_loss": 0.8956291377544403, "train/step_real_loss": 0.8956291377544403, "train/lr": 0.00016722201567174016, "perf/step_duration_sec": 3.435971725033596, "perf/samples_per_sec": 9.31323147011261, "perf/tokens_per_sec": 8253.269313420418, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28358.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
27
+ {"timestamp": 1774538946.1442409, "event": "train_step", "step": 90, "epoch": 1, "metrics": {"train/step_loss": 0.7758428752422333, "train/step_real_loss": 0.7758428752422333, "train/lr": 0.00016666598632252867, "perf/step_duration_sec": 3.4496665820479393, "perf/samples_per_sec": 9.276258803250135, "perf/tokens_per_sec": 8393.564801503364, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28955.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
28
+ {"timestamp": 1774538962.8868124, "event": "eval_step", "step": 90, "epoch": 1, "metrics": {"eval/loss": 0.8181438957867414, "eval/duration_sec": 16.727356909075752}}
29
+ {"timestamp": 1774538980.5067496, "event": "train_step", "step": 95, "epoch": 1, "metrics": {"train/step_loss": 0.9813565611839294, "train/step_real_loss": 0.9813565611839294, "train/lr": 0.000166060427461308, "perf/step_duration_sec": 3.543182759080082, "perf/samples_per_sec": 9.031428005793348, "perf/tokens_per_sec": 7863.551471794196, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27862.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
30
+ {"timestamp": 1774538998.4448798, "event": "train_step", "step": 100, "epoch": 1, "metrics": {"train/step_loss": 0.9215702414512634, "train/step_real_loss": 0.9215702414512634, "train/lr": 0.00016540572651496904, "perf/step_duration_sec": 3.3544326319824904, "perf/samples_per_sec": 9.539616236408898, "perf/tokens_per_sec": 7700.855206841083, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25832.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
31
+ {"timestamp": 1774539015.1820667, "event": "eval_step", "step": 100, "epoch": 1, "metrics": {"eval/loss": 0.8118052698023941, "eval/duration_sec": 16.718859139131382}}
32
+ {"timestamp": 1774539032.8519325, "event": "train_step", "step": 105, "epoch": 1, "metrics": {"train/step_loss": 0.8308639824390411, "train/step_real_loss": 0.8308639824390411, "train/lr": 0.00016470230235072418, "perf/step_duration_sec": 3.6407651680056006, "perf/samples_per_sec": 8.78936117089076, "perf/tokens_per_sec": 6946.891335442784, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25292.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.79835033416748}}
33
+ {"timestamp": 1774539050.0415013, "event": "train_step", "step": 110, "epoch": 1, "metrics": {"train/step_loss": 1.5791892444386202, "train/step_real_loss": 0.8497635126113892, "train/lr": 0.00016395060500812273, "train/step_canary_loss": 13.25, "perf/step_duration_sec": 3.5158363091759384, "perf/samples_per_sec": 9.670529856940101, "perf/tokens_per_sec": 7601.88974960112, "perf/logical_batch_size": 34.0, "perf/logical_token_count": 26727.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.776763916015625, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
34
+ {"timestamp": 1774539066.9950597, "event": "eval_step", "step": 110, "epoch": 1, "metrics": {"eval/loss": 0.806415346491596, "eval/duration_sec": 16.932611752999946}}
35
+ {"timestamp": 1774539084.4092364, "event": "train_step", "step": 115, "epoch": 1, "metrics": {"train/step_loss": 0.8096671402454376, "train/step_real_loss": 0.8096671402454376, "train/lr": 0.00016315111541112274, "perf/step_duration_sec": 3.4490709209349006, "perf/samples_per_sec": 9.277860830801972, "perf/tokens_per_sec": 8118.708093253651, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28002.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
36
+ {"timestamp": 1774539101.9260929, "event": "train_step", "step": 120, "epoch": 1, "metrics": {"train/step_loss": 0.8926524519920349, "train/step_real_loss": 0.8926524519920349, "train/lr": 0.00016230434506040342, "perf/step_duration_sec": 3.5793851560447365, "perf/samples_per_sec": 8.940082892716799, "perf/tokens_per_sec": 7337.014279016518, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26262.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
37
+ {"timestamp": 1774539119.2891507, "event": "eval_step", "step": 120, "epoch": 1, "metrics": {"eval/loss": 0.8018699715966764, "eval/duration_sec": 17.341549766948447}}
38
+ {"timestamp": 1774539136.5355008, "event": "train_step", "step": 125, "epoch": 1, "metrics": {"train/step_loss": 0.8833202421665192, "train/step_real_loss": 0.8833202421665192, "train/lr": 0.00016141083570611557, "perf/step_duration_sec": 3.5714068580418825, "perf/samples_per_sec": 8.960054474875719, "perf/tokens_per_sec": 7589.166140219733, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27104.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
39
+ {"timestamp": 1774539154.3332527, "event": "train_step", "step": 130, "epoch": 1, "metrics": {"train/step_loss": 0.7683468163013458, "train/step_real_loss": 0.7683468163013458, "train/lr": 0.00016047115900127802, "perf/step_duration_sec": 3.4090289890300483, "perf/samples_per_sec": 9.386837161834983, "perf/tokens_per_sec": 7976.758216930584, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27193.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
40
+ {"timestamp": 1774539171.858782, "event": "eval_step", "step": 130, "epoch": 1, "metrics": {"eval/loss": 0.7979619767030944, "eval/duration_sec": 17.51283926097676}}
41
+ {"timestamp": 1774539189.2946777, "event": "train_step", "step": 135, "epoch": 1, "metrics": {"train/step_loss": 0.8694225251674652, "train/step_real_loss": 0.8694225251674652, "train/lr": 0.0001594859161360435, "perf/step_duration_sec": 3.4834307809360325, "perf/samples_per_sec": 9.186345879220049, "perf/tokens_per_sec": 7484.575305094535, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26072.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
42
+ {"timestamp": 1774539207.1762073, "event": "train_step", "step": 140, "epoch": 1, "metrics": {"train/step_loss": 0.8734995722770691, "train/step_real_loss": 0.8734995722770691, "train/lr": 0.0001584557374530666, "perf/step_duration_sec": 3.747924938099459, "perf/samples_per_sec": 8.538057866288787, "perf/tokens_per_sec": 6833.114436114245, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25610.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
43
+ {"timestamp": 1774539223.4354973, "event": "eval_step", "step": 140, "epoch": 1, "metrics": {"eval/loss": 0.795108987585358, "eval/duration_sec": 16.248313741991296}}
44
+ {"timestamp": 1774539241.0867982, "event": "train_step", "step": 145, "epoch": 1, "metrics": {"train/step_loss": 0.8508216738700867, "train/step_real_loss": 0.8508216738700867, "train/lr": 0.0001573812820442207, "perf/step_duration_sec": 3.520246823085472, "perf/samples_per_sec": 9.090271679288733, "perf/tokens_per_sec": 8455.37301668841, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29765.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
45
+ {"timestamp": 1774539258.4430752, "event": "train_step", "step": 150, "epoch": 1, "metrics": {"train/step_loss": 0.747307151556015, "train/step_real_loss": 0.747307151556015, "train/lr": 0.0001562632373289213, "perf/step_duration_sec": 3.5919068260118365, "perf/samples_per_sec": 8.908917059947854, "perf/tokens_per_sec": 7148.570729633783, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25677.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
46
+ {"timestamp": 1774539274.641793, "event": "eval_step", "step": 150, "epoch": 1, "metrics": {"eval/loss": 0.792569605714601, "eval/duration_sec": 16.190975609002635}}
47
+ {"timestamp": 1774539291.9649053, "event": "train_step", "step": 155, "epoch": 1, "metrics": {"train/step_loss": 1.1886012807036892, "train/step_real_loss": 0.8448856770992279, "train/lr": 0.00015510231861432628, "train/step_canary_loss": 12.1875, "perf/step_duration_sec": 3.4058454299811274, "perf/samples_per_sec": 9.689224211265177, "perf/tokens_per_sec": 8437.846223737535, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 28738.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
48
+ {"timestamp": 1774539309.2175198, "event": "train_step", "step": 160, "epoch": 1, "metrics": {"train/step_loss": 1.337690830230713, "train/step_real_loss": 0.8802809417247772, "train/lr": 0.00015389926863769383, "train/step_canary_loss": 8.65625, "perf/step_duration_sec": 3.5819464148953557, "perf/samples_per_sec": 9.492045960992773, "perf/tokens_per_sec": 8081.918780140581, "perf/logical_batch_size": 34.0, "perf/logical_token_count": 28949.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.583967685699463, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
49
+ {"timestamp": 1774539326.0374594, "event": "eval_step", "step": 160, "epoch": 1, "metrics": {"eval/loss": 0.7895505786266015, "eval/duration_sec": 16.80562614183873}}
50
+ {"timestamp": 1774539343.2769709, "event": "train_step", "step": 165, "epoch": 1, "metrics": {"train/step_loss": 0.7841204404830933, "train/step_real_loss": 0.7841204404830933, "train/lr": 0.00015265485709119086, "perf/step_duration_sec": 3.511378558818251, "perf/samples_per_sec": 9.113229879369529, "perf/tokens_per_sec": 7406.492796023855, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26007.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
51
+ {"timestamp": 1774539361.2896898, "event": "train_step", "step": 170, "epoch": 1, "metrics": {"train/step_loss": 0.8921196162700653, "train/step_real_loss": 0.8921196162700653, "train/lr": 0.00015136988012945623, "perf/step_duration_sec": 3.4350620880723, "perf/samples_per_sec": 9.315697701976, "perf/tokens_per_sec": 7599.571515940108, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26105.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
52
+ {"timestamp": 1774539377.4260106, "event": "eval_step", "step": 170, "epoch": 1, "metrics": {"eval/loss": 0.7865845706799757, "eval/duration_sec": 16.12210800917819}}
53
+ {"timestamp": 1774539394.9399402, "event": "train_step", "step": 175, "epoch": 1, "metrics": {"train/step_loss": 0.7969093024730682, "train/step_real_loss": 0.7969093024730682, "train/lr": 0.00015004515986023393, "perf/step_duration_sec": 3.6512177139520645, "perf/samples_per_sec": 8.7641993731903, "perf/tokens_per_sec": 8097.298577136599, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29565.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
54
+ {"timestamp": 1774539412.5012622, "event": "train_step", "step": 180, "epoch": 1, "metrics": {"train/step_loss": 0.9273938238620758, "train/step_real_loss": 0.9273938238620758, "train/lr": 0.00014868154381840137, "perf/step_duration_sec": 3.552449414972216, "perf/samples_per_sec": 9.007869292981974, "perf/tokens_per_sec": 7251.053284935083, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25759.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
55
+ {"timestamp": 1774539428.9990897, "event": "eval_step", "step": 180, "epoch": 1, "metrics": {"eval/loss": 0.7841662333711333, "eval/duration_sec": 16.481489900965244}}
56
+ {"timestamp": 1774539446.471609, "event": "train_step", "step": 185, "epoch": 1, "metrics": {"train/step_loss": 0.8930624723434448, "train/step_real_loss": 0.8930624723434448, "train/lr": 0.0001472799044237304, "perf/step_duration_sec": 3.5634033400565386, "perf/samples_per_sec": 8.9801790440855, "perf/tokens_per_sec": 7711.728754108422, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27480.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
57
+ {"timestamp": 1774539463.785303, "event": "train_step", "step": 190, "epoch": 1, "metrics": {"train/step_loss": 0.7376655340194702, "train/step_real_loss": 0.7376655340194702, "train/lr": 0.00014584113842272677, "perf/step_duration_sec": 3.4315317471046, "perf/samples_per_sec": 9.325281640480354, "perf/tokens_per_sec": 7512.38860656072, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25779.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
58
+ {"timestamp": 1774539480.834944, "event": "eval_step", "step": 190, "epoch": 1, "metrics": {"eval/loss": 0.781792089343071, "eval/duration_sec": 17.031947176903486}}
59
+ {"timestamp": 1774539498.2693355, "event": "train_step", "step": 195, "epoch": 1, "metrics": {"train/step_loss": 0.7903984487056732, "train/step_real_loss": 0.7903984487056732, "train/lr": 0.00014436616631490634, "perf/step_duration_sec": 3.4992341939359903, "perf/samples_per_sec": 9.144858053643425, "perf/tokens_per_sec": 7948.0247558728415, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27812.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
60
+ {"timestamp": 1774539515.6852608, "event": "train_step", "step": 200, "epoch": 1, "metrics": {"train/step_loss": 0.8167151510715485, "train/step_real_loss": 0.8167151510715485, "train/lr": 0.00014285593176387398, "perf/step_duration_sec": 3.357256160117686, "perf/samples_per_sec": 9.53159320404025, "perf/tokens_per_sec": 8705.025355877386, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29225.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
61
+ {"timestamp": 1774539532.5543816, "event": "eval_step", "step": 200, "epoch": 1, "metrics": {"eval/loss": 0.7796081537785737, "eval/duration_sec": 16.85715805203654}}
62
+ {"timestamp": 1774539549.6375906, "event": "train_step", "step": 205, "epoch": 1, "metrics": {"train/step_loss": 0.756825715303421, "train/step_real_loss": 0.756825715303421, "train/lr": 0.00014131140099358272, "perf/step_duration_sec": 3.373895718017593, "perf/samples_per_sec": 9.484584786989892, "perf/tokens_per_sec": 7946.600085124562, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26811.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
63
+ {"timestamp": 1774539566.5938864, "event": "train_step", "step": 210, "epoch": 1, "metrics": {"train/step_loss": 0.8288808166980743, "train/step_real_loss": 0.8288808166980743, "train/lr": 0.00013973356217015901, "perf/step_duration_sec": 3.250324517954141, "perf/samples_per_sec": 9.845170789328394, "perf/tokens_per_sec": 8377.62501854413, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27230.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
64
+ {"timestamp": 1774539582.8153474, "event": "eval_step", "step": 210, "epoch": 1, "metrics": {"eval/loss": 0.7775273114118887, "eval/duration_sec": 16.211750861024484}}
65
+ {"timestamp": 1774539600.2755373, "event": "train_step", "step": 215, "epoch": 1, "metrics": {"train/step_loss": 0.8652437925338745, "train/step_real_loss": 0.8652437925338745, "train/lr": 0.00013812342476969008, "perf/step_duration_sec": 3.554661117028445, "perf/samples_per_sec": 9.002264617210747, "perf/tokens_per_sec": 7815.372291585368, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27781.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.85142803192139}}
66
+ {"timestamp": 1774539627.0015175, "event": "train_epoch", "step": 218, "epoch": 1, "metrics": {"train/epoch_loss": 1.128930080550388, "train/epoch_real_loss": 1.0973064087570532, "train/epoch_canary_loss": 11.634375, "perf/epoch_duration_sec": 1171.5243938071653, "perf/epoch_samples_per_sec": 47.784749763575604, "perf/epoch_tokens_per_sec": 39958.44068416844, "perf/epoch_samples": 55981.0, "perf/epoch_tokens": 46812288.0, "system/cuda_epoch_peak_memory_gb": 97.85142803192139, "eval/loss": 0.7762140348229718, "eval/duration_sec": 16.09541182103567}}
67
+ {"timestamp": 1774539638.0110261, "event": "audit_epoch", "step": 218, "epoch": 1, "metrics": {"audit/delta": 1e-05, "audit/num_canaries": 500.0, "audit/num_members": 250.0, "audit/paper_guess_fraction": 0.2, "audit/paper_guess_steps": 20.0, "audit/loss/auc": 0.762512, "audit/loss/empirical_epsilon/0.05": 3.4791953936219215, "audit/loss/empirical_epsilon/0.01": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/loss/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 100.0, "audit/embedding/auc": 0.75, "audit/embedding/empirical_epsilon/0.05": 3.4791953936219215, "audit/embedding/empirical_epsilon/0.01": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 100.0, "perf/audit_duration_sec": 8.600382861914113}}
68
+ {"timestamp": 1774539645.3794181, "event": "train_step", "step": 220, "epoch": 2, "metrics": {"train/step_loss": 0.8181352019309998, "train/step_real_loss": 0.8181352019309998, "train/lr": 0.00013648201893237688, "perf/step_duration_sec": 3.3851619719062, "perf/samples_per_sec": 9.453018870462099, "perf/tokens_per_sec": 7909.222726179758, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26774.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
69
+ {"timestamp": 1774539661.5299425, "event": "eval_step", "step": 220, "epoch": 2, "metrics": {"eval/loss": 0.7758509523194769, "eval/duration_sec": 16.137880818918347}}
70
+ {"timestamp": 1774539678.3122778, "event": "train_step", "step": 225, "epoch": 2, "metrics": {"train/step_loss": 0.7568798661231995, "train/step_real_loss": 0.7568798661231995, "train/lr": 0.00013481039480346715, "perf/step_duration_sec": 3.338141086976975, "perf/samples_per_sec": 9.586173611666977, "perf/tokens_per_sec": 8389.699317760793, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28006.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
71
+ {"timestamp": 1774539695.119724, "event": "train_step", "step": 230, "epoch": 2, "metrics": {"train/step_loss": 0.8289789855480194, "train/step_real_loss": 0.8289789855480194, "train/lr": 0.00013310962186138903, "perf/step_duration_sec": 3.3264342141337693, "perf/samples_per_sec": 9.619910673126919, "perf/tokens_per_sec": 7569.366588708209, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25179.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
72
+ {"timestamp": 1774539711.8129272, "event": "eval_step", "step": 230, "epoch": 2, "metrics": {"eval/loss": 0.7742470229125542, "eval/duration_sec": 16.68211223790422}}
73
+ {"timestamp": 1774539729.5724974, "event": "train_step", "step": 235, "epoch": 2, "metrics": {"train/step_loss": 0.7927652895450592, "train/step_real_loss": 0.7927652895450592, "train/lr": 0.0001313807882335159, "perf/step_duration_sec": 3.4780818428844213, "perf/samples_per_sec": 9.200473549944402, "perf/tokens_per_sec": 7777.562812485813, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27051.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
74
+ {"timestamp": 1774539747.1401231, "event": "train_step", "step": 240, "epoch": 2, "metrics": {"train/step_loss": 0.7606290280818939, "train/step_real_loss": 0.7606290280818939, "train/lr": 0.00012962500000000005, "perf/step_duration_sec": 3.49273478705436, "perf/samples_per_sec": 9.161875135382834, "perf/tokens_per_sec": 6672.13556734255, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 23304.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
75
+ {"timestamp": 1774539763.9688694, "event": "eval_step", "step": 240, "epoch": 2, "metrics": {"eval/loss": 0.772910956941221, "eval/duration_sec": 16.813530725892633}}
76
+ {"timestamp": 1774539781.2902787, "event": "train_step", "step": 245, "epoch": 2, "metrics": {"train/step_loss": 0.8568435311317444, "train/step_real_loss": 0.8568435311317444, "train/lr": 0.00012784338048611992, "perf/step_duration_sec": 3.5417851228266954, "perf/samples_per_sec": 9.034991929284752, "perf/tokens_per_sec": 7434.386640312401, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26331.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
77
+ {"timestamp": 1774539799.1321223, "event": "train_step", "step": 250, "epoch": 2, "metrics": {"train/step_loss": 0.7391688227653503, "train/step_real_loss": 0.7391688227653503, "train/lr": 0.00012603706954359485, "perf/step_duration_sec": 3.6043081579264253, "perf/samples_per_sec": 8.878264176615172, "perf/tokens_per_sec": 7267.414120070556, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26194.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
78
+ {"timestamp": 1774539815.0383542, "event": "eval_step", "step": 250, "epoch": 2, "metrics": {"eval/loss": 0.771633688684391, "eval/duration_sec": 15.891243164194748}}
79
+ {"timestamp": 1774539832.655197, "event": "train_step", "step": 255, "epoch": 2, "metrics": {"train/step_loss": 0.7805173993110657, "train/step_real_loss": 0.7805173993110657, "train/lr": 0.0001242072228213259, "perf/step_duration_sec": 3.611758361104876, "perf/samples_per_sec": 8.85995041767159, "perf/tokens_per_sec": 7061.934229785361, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25506.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
80
+ {"timestamp": 1774539849.396755, "event": "train_step", "step": 260, "epoch": 2, "metrics": {"train/step_loss": 0.758133053779602, "train/step_real_loss": 0.758133053779602, "train/lr": 0.00012235501102603023, "perf/step_duration_sec": 3.3163676189724356, "perf/samples_per_sec": 9.649111219435644, "perf/tokens_per_sec": 8388.696066396862, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27820.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
81
+ {"timestamp": 1774539866.0030868, "event": "eval_step", "step": 260, "epoch": 2, "metrics": {"eval/loss": 0.7709154452642668, "eval/duration_sec": 16.590645326999947}}
82
+ {"timestamp": 1774539884.1196933, "event": "train_step", "step": 265, "epoch": 2, "metrics": {"train/step_loss": 0.940812739458951, "train/step_real_loss": 0.7827131450176239, "train/lr": 0.0001204816191732417, "train/step_canary_loss": 6.0, "perf/step_duration_sec": 3.640273388940841, "perf/samples_per_sec": 9.065253203304476, "perf/tokens_per_sec": 7517.2925426796055, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 27365.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
83
+ {"timestamp": 1774539904.1036265, "event": "train_step", "step": 270, "epoch": 2, "metrics": {"train/step_loss": 0.8124397099018097, "train/step_real_loss": 0.8124397099018097, "train/lr": 0.00011858824582915682, "perf/step_duration_sec": 3.5830548191443086, "perf/samples_per_sec": 8.93092671343558, "perf/tokens_per_sec": 7556.959456865476, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27077.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
84
+ {"timestamp": 1774539921.216135, "event": "eval_step", "step": 270, "epoch": 2, "metrics": {"eval/loss": 0.7695811219189478, "eval/duration_sec": 17.09733649296686}}
85
+ {"timestamp": 1774539939.995836, "event": "train_step", "step": 275, "epoch": 2, "metrics": {"train/step_loss": 0.7777769863605499, "train/step_real_loss": 0.7777769863605499, "train/lr": 0.00011667610234381102, "perf/step_duration_sec": 3.4810175739694387, "perf/samples_per_sec": 9.19271429115771, "perf/tokens_per_sec": 8018.345040462313, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27912.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
86
+ {"timestamp": 1774539959.6496387, "event": "train_step", "step": 280, "epoch": 2, "metrics": {"train/step_loss": 0.8779713809490204, "train/step_real_loss": 0.8779713809490204, "train/lr": 0.00011474641207607662, "perf/step_duration_sec": 3.495331821963191, "perf/samples_per_sec": 9.155067853336698, "perf/tokens_per_sec": 6662.600630265782, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 23288.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
87
+ {"timestamp": 1774539975.7933335, "event": "eval_step", "step": 280, "epoch": 2, "metrics": {"eval/loss": 0.7681934292549673, "eval/duration_sec": 16.12853733310476}}
88
+ {"timestamp": 1774539994.1960711, "event": "train_step", "step": 285, "epoch": 2, "metrics": {"train/step_loss": 0.8127053678035736, "train/step_real_loss": 0.8127053678035736, "train/lr": 0.000112800409610977, "perf/step_duration_sec": 3.3115670438855886, "perf/samples_per_sec": 9.663098942563812, "perf/tokens_per_sec": 8591.400875465095, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28451.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
89
+ {"timestamp": 1774540011.8948312, "event": "train_step", "step": 290, "epoch": 2, "metrics": {"train/step_loss": 0.8134279549121857, "train/step_real_loss": 0.8134279549121857, "train/lr": 0.00011083933996981912, "perf/step_duration_sec": 3.478434829041362, "perf/samples_per_sec": 9.199539900196731, "perf/tokens_per_sec": 7699.727410842782, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26783.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
90
+ {"timestamp": 1774540028.3504686, "event": "eval_step", "step": 290, "epoch": 2, "metrics": {"eval/loss": 0.7673408377753652, "eval/duration_sec": 16.421336075058207}}
91
+ {"timestamp": 1774540045.6837075, "event": "train_step", "step": 295, "epoch": 2, "metrics": {"train/step_loss": 0.6896935403347015, "train/step_real_loss": 0.6896935403347015, "train/lr": 0.00010886445781364846, "perf/step_duration_sec": 3.4652010852005333, "perf/samples_per_sec": 9.234673317132515, "perf/tokens_per_sec": 8122.760933041528, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28147.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
92
+ {"timestamp": 1774540062.6908014, "event": "train_step", "step": 300, "epoch": 2, "metrics": {"train/step_loss": 0.7095499038696289, "train/step_real_loss": 0.7095499038696289, "train/lr": 0.00010687702664053723, "perf/step_duration_sec": 3.5274920598603785, "perf/samples_per_sec": 9.07160085890217, "perf/tokens_per_sec": 7587.260168364303, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26764.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
93
+ {"timestamp": 1774540079.110272, "event": "eval_step", "step": 300, "epoch": 2, "metrics": {"eval/loss": 0.7664490196367968, "eval/duration_sec": 16.40780350100249}}
94
+ {"timestamp": 1774540096.5141547, "event": "train_step", "step": 305, "epoch": 2, "metrics": {"train/step_loss": 0.7463825941085815, "train/step_real_loss": 0.7463825941085815, "train/lr": 0.0001048783179772182, "perf/step_duration_sec": 3.499851844040677, "perf/samples_per_sec": 9.143244178889328, "perf/tokens_per_sec": 7616.6081273954005, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26657.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
95
+ {"timestamp": 1774540113.608032, "event": "train_step", "step": 310, "epoch": 2, "metrics": {"train/step_loss": 0.8096683919429779, "train/step_real_loss": 0.8096683919429779, "train/lr": 0.00010286961056558237, "perf/step_duration_sec": 3.3904766468331218, "perf/samples_per_sec": 9.4382009768124, "perf/tokens_per_sec": 7991.206789554849, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27094.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
96
+ {"timestamp": 1774540129.9482985, "event": "eval_step", "step": 310, "epoch": 2, "metrics": {"eval/loss": 0.7657463623453742, "eval/duration_sec": 16.329357700888067}}
97
+ {"timestamp": 1774540147.0143843, "event": "train_step", "step": 315, "epoch": 2, "metrics": {"train/step_loss": 0.7295234203338623, "train/step_real_loss": 0.7295234203338623, "train/lr": 0.00010085218954456027, "perf/step_duration_sec": 3.2363929860293865, "perf/samples_per_sec": 9.887550782038877, "perf/tokens_per_sec": 7940.321249901096, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25698.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
98
+ {"timestamp": 1774540164.381428, "event": "train_step", "step": 320, "epoch": 2, "metrics": {"train/step_loss": 0.7396710515022278, "train/step_real_loss": 0.7396710515022278, "train/lr": 9.882734562791052e-05, "perf/step_duration_sec": 3.4866965510882437, "perf/samples_per_sec": 9.177741604732532, "perf/tokens_per_sec": 8442.661863078485, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29437.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
99
+ {"timestamp": 1774540180.8920996, "event": "eval_step", "step": 320, "epoch": 2, "metrics": {"eval/loss": 0.7646235585536646, "eval/duration_sec": 16.49111690209247}}
100
+ {"timestamp": 1774540198.700798, "event": "train_step", "step": 325, "epoch": 2, "metrics": {"train/step_loss": 0.7767525017261505, "train/step_real_loss": 0.7767525017261505, "train/lr": 9.679637427844199e-05, "perf/step_duration_sec": 3.5239353249780834, "perf/samples_per_sec": 9.080756895048584, "perf/tokens_per_sec": 8106.845718054623, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28568.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
101
+ {"timestamp": 1774540215.8494503, "event": "train_step", "step": 330, "epoch": 2, "metrics": {"train/step_loss": 0.8057853579521179, "train/step_real_loss": 0.8057853579521179, "train/lr": 9.476057487919712e-05, "perf/step_duration_sec": 3.3903933928813785, "perf/samples_per_sec": 9.438432739748912, "perf/tokens_per_sec": 6803.045348197145, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 23065.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
102
+ {"timestamp": 1774540232.5870273, "event": "eval_step", "step": 330, "epoch": 2, "metrics": {"eval/loss": 0.7636887612550155, "eval/duration_sec": 16.72419739002362}}
103
+ {"timestamp": 1774540249.7668092, "event": "train_step", "step": 335, "epoch": 2, "metrics": {"train/step_loss": 0.7550980150699615, "train/step_real_loss": 0.7550980150699615, "train/lr": 9.272124990212752e-05, "perf/step_duration_sec": 3.5105914149899036, "perf/samples_per_sec": 9.115273245232395, "perf/tokens_per_sec": 6794.011942876182, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 23851.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
104
+ {"timestamp": 1774540266.7018678, "event": "train_step", "step": 340, "epoch": 2, "metrics": {"train/step_loss": 0.7399545907974243, "train/step_real_loss": 0.7399545907974243, "train/lr": 9.067970407479325e-05, "perf/step_duration_sec": 3.379189013969153, "perf/samples_per_sec": 9.46972775648711, "perf/tokens_per_sec": 7692.970086176217, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25996.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
105
+ {"timestamp": 1774540283.7666714, "event": "eval_step", "step": 340, "epoch": 2, "metrics": {"eval/loss": 0.7627839732753194, "eval/duration_sec": 17.052499413955957}}
106
+ {"timestamp": 1774540300.6349876, "event": "train_step", "step": 345, "epoch": 2, "metrics": {"train/step_loss": 0.7904732823371887, "train/step_real_loss": 0.7904732823371887, "train/lr": 8.863724354561888e-05, "perf/step_duration_sec": 3.4896194499451667, "perf/samples_per_sec": 9.170054345181628, "perf/tokens_per_sec": 7245.489189486634, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25284.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
107
+ {"timestamp": 1774540318.267128, "event": "train_step", "step": 350, "epoch": 2, "metrics": {"train/step_loss": 0.8455444872379303, "train/step_real_loss": 0.8455444872379303, "train/lr": 8.659517504824063e-05, "perf/step_duration_sec": 3.494046765845269, "perf/samples_per_sec": 9.158434945062522, "perf/tokens_per_sec": 7033.105635623951, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 24574.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
108
+ {"timestamp": 1774540334.801836, "event": "eval_step", "step": 350, "epoch": 2, "metrics": {"eval/loss": 0.7622423175236454, "eval/duration_sec": 16.52213899488561}}
109
+ {"timestamp": 1774540352.5217278, "event": "train_step", "step": 355, "epoch": 2, "metrics": {"train/step_loss": 0.774720162153244, "train/step_real_loss": 0.774720162153244, "train/lr": 8.455480506547902e-05, "perf/step_duration_sec": 3.439865805907175, "perf/samples_per_sec": 9.302688478442208, "perf/tokens_per_sec": 8260.20595082678, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28414.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
110
+ {"timestamp": 1774540369.9482565, "event": "train_step", "step": 360, "epoch": 2, "metrics": {"train/step_loss": 0.7312933206558228, "train/step_real_loss": 0.7312933206558228, "train/lr": 8.251743899347202e-05, "perf/step_duration_sec": 3.391233369940892, "perf/samples_per_sec": 9.436094927479955, "perf/tokens_per_sec": 6997.749022625837, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 23731.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
111
+ {"timestamp": 1774540386.0299351, "event": "eval_step", "step": 360, "epoch": 2, "metrics": {"eval/loss": 0.7614095991720324, "eval/duration_sec": 16.063601257978007}}
112
+ {"timestamp": 1774540403.7036886, "event": "train_step", "step": 365, "epoch": 2, "metrics": {"train/step_loss": 0.8506239652633667, "train/step_real_loss": 0.8506239652633667, "train/lr": 8.048438030650353e-05, "perf/step_duration_sec": 3.4739691771101207, "perf/samples_per_sec": 9.211365550059295, "perf/tokens_per_sec": 7682.85457909633, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26690.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
113
+ {"timestamp": 1774540421.42377, "event": "train_step", "step": 370, "epoch": 2, "metrics": {"train/step_loss": 0.6848389506340027, "train/step_real_loss": 0.6848389506340027, "train/lr": 7.845692972306124e-05, "perf/step_duration_sec": 3.5644277748651803, "perf/samples_per_sec": 8.977598094608709, "perf/tokens_per_sec": 7101.560642775945, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25313.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
114
+ {"timestamp": 1774540438.2559083, "event": "eval_step", "step": 370, "epoch": 2, "metrics": {"eval/loss": 0.7608118504285812, "eval/duration_sec": 16.818945176899433}}
115
+ {"timestamp": 1774540456.1543565, "event": "train_step", "step": 375, "epoch": 2, "metrics": {"train/step_loss": 0.7430407106876373, "train/step_real_loss": 0.7430407106876373, "train/lr": 7.643638437365766e-05, "perf/step_duration_sec": 3.5129232809413224, "perf/samples_per_sec": 9.1092225593453, "perf/tokens_per_sec": 8359.41967742919, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29366.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
116
+ {"timestamp": 1774540473.868304, "event": "train_step", "step": 380, "epoch": 2, "metrics": {"train/step_loss": 0.814172238111496, "train/step_real_loss": 0.814172238111496, "train/lr": 7.442403697094692e-05, "perf/step_duration_sec": 3.501377285923809, "perf/samples_per_sec": 9.139260749947166, "perf/tokens_per_sec": 7028.948322404678, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 24611.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
117
+ {"timestamp": 1774540490.6731074, "event": "eval_step", "step": 380, "epoch": 2, "metrics": {"eval/loss": 0.7602399059611818, "eval/duration_sec": 16.789929745020345}}
118
+ {"timestamp": 1774540507.9563136, "event": "train_step", "step": 385, "epoch": 2, "metrics": {"train/step_loss": 0.7185265719890594, "train/step_real_loss": 0.7185265719890594, "train/lr": 7.242117498266774e-05, "perf/step_duration_sec": 3.3475870778784156, "perf/samples_per_sec": 9.559124006500971, "perf/tokens_per_sec": 8158.413616923376, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27311.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
119
+ {"timestamp": 1774540525.4224126, "event": "train_step", "step": 390, "epoch": 2, "metrics": {"train/step_loss": 0.7320943176746368, "train/step_real_loss": 0.7320943176746368, "train/lr": 7.042907980794205e-05, "perf/step_duration_sec": 3.506713897921145, "perf/samples_per_sec": 9.125352375901064, "perf/tokens_per_sec": 7523.56786666868, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26383.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
120
+ {"timestamp": 1774540542.4360893, "event": "eval_step", "step": 390, "epoch": 2, "metrics": {"eval/loss": 0.7598179606315882, "eval/duration_sec": 16.986541678197682}}
121
+ {"timestamp": 1774540559.646183, "event": "train_step", "step": 395, "epoch": 2, "metrics": {"train/step_loss": 0.8207666277885437, "train/step_real_loss": 0.8207666277885437, "train/lr": 6.844902595745646e-05, "perf/step_duration_sec": 3.3661333608906716, "perf/samples_per_sec": 9.50645639052544, "perf/tokens_per_sec": 7737.364271601095, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26045.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
122
+ {"timestamp": 1774540577.133158, "event": "train_step", "step": 400, "epoch": 2, "metrics": {"train/step_loss": 0.7455478608608246, "train/step_real_loss": 0.7455478608608246, "train/lr": 6.648228023805054e-05, "perf/step_duration_sec": 3.465459569124505, "perf/samples_per_sec": 9.233984515388332, "perf/tokens_per_sec": 7644.873492693221, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26493.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
123
+ {"timestamp": 1774540593.9303052, "event": "eval_step", "step": 400, "epoch": 2, "metrics": {"eval/loss": 0.7594009866209134, "eval/duration_sec": 16.78406825219281}}
124
+ {"timestamp": 1774540611.4112008, "event": "train_step", "step": 405, "epoch": 2, "metrics": {"train/step_loss": 0.7589297294616699, "train/step_real_loss": 0.7589297294616699, "train/lr": 6.453010094223415e-05, "perf/step_duration_sec": 3.5417204569093883, "perf/samples_per_sec": 9.035156893190877, "perf/tokens_per_sec": 8226.228002597381, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29135.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
125
+ {"timestamp": 1774540628.3848057, "event": "train_step", "step": 410, "epoch": 2, "metrics": {"train/step_loss": 0.7410624921321869, "train/step_real_loss": 0.7410624921321869, "train/lr": 6.259373704315176e-05, "perf/step_duration_sec": 3.3305229651741683, "perf/samples_per_sec": 9.608100690074831, "perf/tokens_per_sec": 7923.380284699522, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26389.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
126
+ {"timestamp": 1774540644.9026184, "event": "eval_step", "step": 410, "epoch": 2, "metrics": {"eval/loss": 0.7589393964280253, "eval/duration_sec": 16.50291042099707}}
127
+ {"timestamp": 1774540662.1578808, "event": "train_step", "step": 415, "epoch": 2, "metrics": {"train/step_loss": 0.7675952613353729, "train/step_real_loss": 0.7675952613353729, "train/lr": 6.067442739550964e-05, "perf/step_duration_sec": 3.5111529449932277, "perf/samples_per_sec": 9.113815462134966, "perf/tokens_per_sec": 8433.696983273145, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29612.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
128
+ {"timestamp": 1774540679.7656848, "event": "train_step", "step": 420, "epoch": 2, "metrics": {"train/step_loss": 0.7931506037712097, "train/step_real_loss": 0.7931506037712097, "train/lr": 5.8773399942976274e-05, "perf/step_duration_sec": 3.4180779478047043, "perf/samples_per_sec": 9.36198661606074, "perf/tokens_per_sec": 7714.569533715801, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26369.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
129
+ {"timestamp": 1774540696.4696214, "event": "eval_step", "step": 420, "epoch": 2, "metrics": {"eval/loss": 0.7585963477259098, "eval/duration_sec": 16.686920657986775}}
130
+ {"timestamp": 1774540713.349685, "event": "train_step", "step": 425, "epoch": 2, "metrics": {"train/step_loss": 0.7320750057697296, "train/step_real_loss": 0.7320750057697296, "train/lr": 5.689187093256359e-05, "perf/step_duration_sec": 3.37516968511045, "perf/samples_per_sec": 9.481004804341511, "perf/tokens_per_sec": 9043.100894940988, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 30522.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
131
+ {"timestamp": 1774540730.586966, "event": "train_step", "step": 430, "epoch": 2, "metrics": {"train/step_loss": 0.7799577414989471, "train/step_real_loss": 0.7799577414989471, "train/lr": 5.503104413649151e-05, "perf/step_duration_sec": 3.321356100961566, "perf/samples_per_sec": 9.634618820528061, "perf/tokens_per_sec": 7454.485230545445, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 24759.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
132
+ {"timestamp": 1774540746.7284963, "event": "eval_step", "step": 430, "epoch": 2, "metrics": {"eval/loss": 0.7584100974642711, "eval/duration_sec": 16.1323861151468}}
133
+ {"timestamp": 1774540764.1787012, "event": "train_step", "step": 435, "epoch": 2, "metrics": {"train/step_loss": 0.7374658584594727, "train/step_real_loss": 0.7374658584594727, "train/lr": 5.3192110082033856e-05, "perf/step_duration_sec": 3.5796759920194745, "perf/samples_per_sec": 8.93935654269849, "perf/tokens_per_sec": 7033.318114859997, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25177.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 97.27211284637451}}
134
+ {"timestamp": 1774540784.6702416, "event": "train_epoch", "step": 436, "epoch": 2, "metrics": {"train/epoch_loss": 0.8010046102645352, "train/epoch_real_loss": 0.7772409853038438, "train/epoch_canary_loss": 8.71698641877637, "perf/epoch_duration_sec": 1130.1809665749315, "perf/epoch_samples_per_sec": 49.533660232888344, "perf/epoch_tokens_per_sec": 41420.05341132803, "perf/epoch_samples": 55982.0, "perf/epoch_tokens": 46812156.0, "system/cuda_epoch_peak_memory_gb": 97.27211284637451, "eval/loss": 0.7581099835426912, "eval/duration_sec": 16.36720763379708}}
135
+ {"timestamp": 1774540795.2131171, "event": "audit_epoch", "step": 436, "epoch": 2, "metrics": {"audit/delta": 1e-05, "audit/num_canaries": 500.0, "audit/num_members": 250.0, "audit/paper_guess_fraction": 0.2, "audit/paper_guess_steps": 20.0, "audit/loss/auc": 0.88128, "audit/loss/empirical_epsilon/0.05": 3.4791953936219215, "audit/loss/empirical_epsilon/0.01": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/loss/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 100.0, "audit/embedding/auc": 0.81, "audit/embedding/empirical_epsilon/0.05": 3.4791953936219215, "audit/embedding/empirical_epsilon/0.01": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 100.0, "perf/audit_duration_sec": 8.23527004988864}}
136
+ {"timestamp": 1774540809.4862168, "event": "train_step", "step": 440, "epoch": 3, "metrics": {"train/step_loss": 0.8157035112380981, "train/step_real_loss": 0.8157035112380981, "train/lr": 5.137624528983785e-05, "perf/step_duration_sec": 3.3810758979525417, "perf/samples_per_sec": 9.46444296603281, "perf/tokens_per_sec": 7494.6557737272315, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25340.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 89.7595157623291}}
137
+ {"timestamp": 1774540826.5334082, "event": "eval_step", "step": 440, "epoch": 3, "metrics": {"eval/loss": 0.7580130472779274, "eval/duration_sec": 17.036596653982997}}
138
+ {"timestamp": 1774540843.624869, "event": "train_step", "step": 445, "epoch": 3, "metrics": {"train/step_loss": 0.808228611946106, "train/step_real_loss": 0.808228611946106, "train/lr": 4.958461152120504e-05, "perf/step_duration_sec": 3.451165011851117, "perf/samples_per_sec": 9.27223122919759, "perf/tokens_per_sec": 8102.481308189132, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27963.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.21671152114868}}
139
+ {"timestamp": 1774540861.1815977, "event": "train_step", "step": 450, "epoch": 3, "metrics": {"train/step_loss": 0.7727910876274109, "train/step_real_loss": 0.7727910876274109, "train/lr": 4.781835503481501e-05, "perf/step_duration_sec": 3.66859296313487, "perf/samples_per_sec": 8.722690230713276, "perf/tokens_per_sec": 7021.220467584768, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25758.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.21671152114868}}
140
+ {"timestamp": 1774540877.6283762, "event": "eval_step", "step": 450, "epoch": 3, "metrics": {"eval/loss": 0.7579258727314678, "eval/duration_sec": 16.43751646298915}}
141
+ {"timestamp": 1774540895.1775277, "event": "train_step", "step": 455, "epoch": 3, "metrics": {"train/step_loss": 1.0197869900501135, "train/step_real_loss": 0.6981397271156311, "train/lr": 4.607860585336731e-05, "train/step_canary_loss": 11.3125, "perf/step_duration_sec": 3.599183855112642, "perf/samples_per_sec": 9.168745284607644, "perf/tokens_per_sec": 7657.291516478384, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 27560.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.583967685699463, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
142
+ {"timestamp": 1774540912.624193, "event": "train_step", "step": 460, "epoch": 3, "metrics": {"train/step_loss": 0.7413330674171448, "train/step_real_loss": 0.7413330674171448, "train/lr": 4.4366477040611056e-05, "perf/step_duration_sec": 3.6398022188805044, "perf/samples_per_sec": 8.79168649164741, "perf/tokens_per_sec": 7659.207375442078, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27878.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
143
+ {"timestamp": 1774540928.8435848, "event": "eval_step", "step": 460, "epoch": 3, "metrics": {"eval/loss": 0.7578229023062665, "eval/duration_sec": 16.20310543803498}}
144
+ {"timestamp": 1774540946.0828743, "event": "train_step", "step": 465, "epoch": 3, "metrics": {"train/step_loss": 0.7863788306713104, "train/step_real_loss": 0.7863788306713104, "train/lr": 4.268306398922441e-05, "perf/step_duration_sec": 3.399363676086068, "perf/samples_per_sec": 9.413526485887472, "perf/tokens_per_sec": 6305.297709328502, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 21434.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
145
+ {"timestamp": 1774540963.2547693, "event": "train_step", "step": 470, "epoch": 3, "metrics": {"train/step_loss": 0.6813286416458361, "train/step_real_loss": 0.6576982736587524, "train/lr": 4.102944371999997e-05, "train/step_canary_loss": 1.4375, "perf/step_duration_sec": 3.3317034628707916, "perf/samples_per_sec": 9.9048430833533, "perf/tokens_per_sec": 9127.462974690117, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 30410.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.583967685699463, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
146
+ {"timestamp": 1774540979.9091241, "event": "eval_step", "step": 470, "epoch": 3, "metrics": {"eval/loss": 0.7576829930362494, "eval/duration_sec": 16.62321908492595}}
147
+ {"timestamp": 1774540996.9796317, "event": "train_step", "step": 475, "epoch": 3, "metrics": {"train/step_loss": 0.8033013641834259, "train/step_real_loss": 0.8033013641834259, "train/lr": 3.9406674192783864e-05, "perf/step_duration_sec": 3.3658552318811417, "perf/samples_per_sec": 9.507241932718399, "perf/tokens_per_sec": 7670.858733151761, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25819.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
148
+ {"timestamp": 1774541013.8382132, "event": "train_step", "step": 480, "epoch": 3, "metrics": {"train/step_loss": 0.7144083678722382, "train/step_real_loss": 0.7144083678722382, "train/lr": 3.7815793629610214e-05, "perf/step_duration_sec": 3.285923562012613, "perf/samples_per_sec": 9.738510161934549, "perf/tokens_per_sec": 8122.830460381095, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26691.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
149
+ {"timestamp": 1774541029.969959, "event": "eval_step", "step": 480, "epoch": 3, "metrics": {"eval/loss": 0.7575410879824472, "eval/duration_sec": 16.117679562885314}}
150
+ {"timestamp": 1774541046.9944181, "event": "train_step", "step": 485, "epoch": 3, "metrics": {"train/step_loss": 0.83707594871521, "train/step_real_loss": 0.83707594871521, "train/lr": 3.62578198504629e-05, "perf/step_duration_sec": 3.4277625288814306, "perf/samples_per_sec": 9.33553585768453, "perf/tokens_per_sec": 7473.971660563124, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25619.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
151
+ {"timestamp": 1774541064.0983455, "event": "train_step", "step": 490, "epoch": 3, "metrics": {"train/step_loss": 0.7586729526519775, "train/step_real_loss": 0.7586729526519775, "train/lr": 3.473374962209096e-05, "perf/step_duration_sec": 3.4081453951075673, "perf/samples_per_sec": 9.389270788134912, "perf/tokens_per_sec": 7295.169987668697, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 24863.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
152
+ {"timestamp": 1774541080.9186583, "event": "eval_step", "step": 490, "epoch": 3, "metrics": {"eval/loss": 0.7574281342651533, "eval/duration_sec": 16.801857060054317}}
153
+ {"timestamp": 1774541098.308155, "event": "train_step", "step": 495, "epoch": 3, "metrics": {"train/step_loss": 0.7384476959705353, "train/step_real_loss": 0.7384476959705353, "train/lr": 3.324455802029283e-05, "perf/step_duration_sec": 3.482939150882885, "perf/samples_per_sec": 9.187642566735159, "perf/tokens_per_sec": 7615.120118672394, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26523.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
154
+ {"timestamp": 1774541115.811306, "event": "train_step", "step": 500, "epoch": 3, "metrics": {"train/step_loss": 0.730824887752533, "train/step_real_loss": 0.730824887752533, "train/lr": 3.179119780607868e-05, "perf/step_duration_sec": 3.309445995138958, "perf/samples_per_sec": 9.669292095112848, "perf/tokens_per_sec": 8802.379625710386, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29131.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
155
+ {"timestamp": 1774541132.2387252, "event": "eval_step", "step": 500, "epoch": 3, "metrics": {"eval/loss": 0.757328099530676, "eval/duration_sec": 16.411933867027983}}
156
+ {"timestamp": 1774541149.6690612, "event": "train_step", "step": 505, "epoch": 3, "metrics": {"train/step_loss": 0.7649378180503845, "train/step_real_loss": 0.7649378180503845, "train/lr": 3.037459881610879e-05, "perf/step_duration_sec": 3.5057794600725174, "perf/samples_per_sec": 9.12778466656259, "perf/tokens_per_sec": 7529.566620101647, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26397.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
157
+ {"timestamp": 1774541167.108118, "event": "train_step", "step": 510, "epoch": 3, "metrics": {"train/step_loss": 0.9982280911821307, "train/step_real_loss": 0.8175086379051208, "train/lr": 2.899566736779931e-05, "train/step_canary_loss": 6.78125, "perf/step_duration_sec": 3.4239062799606472, "perf/samples_per_sec": 9.638114277000387, "perf/tokens_per_sec": 8121.133502614327, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 27806.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
158
+ {"timestamp": 1774541183.091979, "event": "eval_step", "step": 510, "epoch": 3, "metrics": {"eval/loss": 0.757218228896027, "eval/duration_sec": 15.971923864912242}}
159
+ {"timestamp": 1774541200.2976594, "event": "train_step", "step": 515, "epoch": 3, "metrics": {"train/step_loss": 0.7422693371772766, "train/step_real_loss": 0.7422693371772766, "train/lr": 2.7655285679474514e-05, "perf/step_duration_sec": 3.629742390010506, "perf/samples_per_sec": 8.816052645517738, "perf/tokens_per_sec": 7802.757594573543, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28322.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
160
+ {"timestamp": 1774541217.702612, "event": "train_step", "step": 520, "epoch": 3, "metrics": {"train/step_loss": 0.7527430951595306, "train/step_real_loss": 0.7527430951595306, "train/lr": 2.635431130593784e-05, "perf/step_duration_sec": 3.4493347899988294, "perf/samples_per_sec": 9.277151088024963, "perf/tokens_per_sec": 7784.399495767447, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26851.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
161
+ {"timestamp": 1774541234.222328, "event": "eval_step", "step": 520, "epoch": 3, "metrics": {"eval/loss": 0.7570996363849744, "eval/duration_sec": 16.504829068901017}}
162
+ {"timestamp": 1774541251.1504445, "event": "train_step", "step": 525, "epoch": 3, "metrics": {"train/step_loss": 0.6985901892185211, "train/step_real_loss": 0.6985901892185211, "train/lr": 2.5093576589821744e-05, "perf/step_duration_sec": 3.4935576680582017, "perf/samples_per_sec": 9.159717125203867, "perf/tokens_per_sec": 7901.687226289149, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27605.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
163
+ {"timestamp": 1774541268.2331986, "event": "train_step", "step": 530, "epoch": 3, "metrics": {"train/step_loss": 0.7169545888900757, "train/step_real_loss": 0.7169545888900757, "train/lr": 2.3873888129068525e-05, "perf/step_duration_sec": 3.2948195259086788, "perf/samples_per_sec": 9.712216328806269, "perf/tokens_per_sec": 8211.375399245426, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27055.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
164
+ {"timestamp": 1774541284.8088741, "event": "eval_step", "step": 530, "epoch": 3, "metrics": {"eval/loss": 0.7570720648635989, "eval/duration_sec": 16.546695807017386}}
165
+ {"timestamp": 1774541301.4996936, "event": "train_step", "step": 535, "epoch": 3, "metrics": {"train/step_loss": 0.766697883605957, "train/step_real_loss": 0.766697883605957, "train/lr": 2.269602626088155e-05, "perf/step_duration_sec": 3.428312035044655, "perf/samples_per_sec": 9.334039513583305, "perf/tokens_per_sec": 7028.823442463027, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 24097.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
166
+ {"timestamp": 1774541318.3984377, "event": "train_step", "step": 540, "epoch": 3, "metrics": {"train/step_loss": 0.7419108152389526, "train/step_real_loss": 0.7419108152389526, "train/lr": 2.156074456247808e-05, "perf/step_duration_sec": 3.3373393951915205, "perf/samples_per_sec": 9.588476391135403, "perf/tokens_per_sec": 8093.57299377995, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27011.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
167
+ {"timestamp": 1774541335.1209495, "event": "eval_step", "step": 540, "epoch": 3, "metrics": {"eval/loss": 0.7569567442912122, "eval/duration_sec": 16.714531823061407}}
168
+ {"timestamp": 1774541352.172168, "event": "train_step", "step": 545, "epoch": 3, "metrics": {"train/step_loss": 0.8203366696834564, "train/step_real_loss": 0.8203366696834564, "train/lr": 2.046876936896261e-05, "perf/step_duration_sec": 3.486993705155328, "perf/samples_per_sec": 9.176959497428907, "perf/tokens_per_sec": 8058.230778695464, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28099.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
169
+ {"timestamp": 1774541369.4427552, "event": "train_step", "step": 550, "epoch": 3, "metrics": {"train/step_loss": 0.8087502158049381, "train/step_real_loss": 0.7080471217632294, "train/lr": 1.9420799308629227e-05, "train/step_canary_loss": 4.03125, "perf/step_duration_sec": 3.471604100894183, "perf/samples_per_sec": 9.505692193271742, "perf/tokens_per_sec": 7816.559495655212, "perf/logical_batch_size": 33.0, "perf/logical_token_count": 27136.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
170
+ {"timestamp": 1774541385.9584184, "event": "eval_step", "step": 550, "epoch": 3, "metrics": {"eval/loss": 0.7570259718467359, "eval/duration_sec": 16.504026405978948}}
171
+ {"timestamp": 1774541404.5539038, "event": "train_step", "step": 555, "epoch": 3, "metrics": {"train/step_loss": 0.8542532324790955, "train/step_real_loss": 0.8542532324790955, "train/lr": 1.8417504855990475e-05, "perf/step_duration_sec": 3.5637081519234926, "perf/samples_per_sec": 8.979410949442695, "perf/tokens_per_sec": 8409.779567337424, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 29970.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
172
+ {"timestamp": 1774541421.8556101, "event": "train_step", "step": 560, "epoch": 3, "metrics": {"train/step_loss": 0.7152487337589264, "train/step_real_loss": 0.7152487337589264, "train/lr": 1.7459527902818278e-05, "perf/step_duration_sec": 3.424267982831225, "perf/samples_per_sec": 9.345062991694366, "perf/tokens_per_sec": 8087.567952874493, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27694.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
173
+ {"timestamp": 1774541437.678642, "event": "eval_step", "step": 560, "epoch": 3, "metrics": {"eval/loss": 0.7569553168571515, "eval/duration_sec": 15.808741455897689}}
174
+ {"timestamp": 1774541455.0345526, "event": "train_step", "step": 565, "epoch": 3, "metrics": {"train/step_loss": 0.7465071380138397, "train/step_real_loss": 0.7465071380138397, "train/lr": 1.654748134747205e-05, "perf/step_duration_sec": 3.620251423912123, "perf/samples_per_sec": 8.839165089097625, "perf/tokens_per_sec": 7200.328636797118, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26067.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
175
+ {"timestamp": 1774541472.0630367, "event": "train_step", "step": 570, "epoch": 3, "metrics": {"train/step_loss": 0.7743452787399292, "train/step_real_loss": 0.7743452787399292, "train/lr": 1.5681948702775953e-05, "perf/step_duration_sec": 3.435435759834945, "perf/samples_per_sec": 9.314684435123139, "perf/tokens_per_sec": 7450.292128655524, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25595.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
176
+ {"timestamp": 1774541488.7719223, "event": "eval_step", "step": 570, "epoch": 3, "metrics": {"eval/loss": 0.7569032507422178, "eval/duration_sec": 16.699837388005108}}
177
+ {"timestamp": 1774541505.6721282, "event": "train_step", "step": 575, "epoch": 3, "metrics": {"train/step_loss": 0.7983605563640594, "train/step_real_loss": 0.7983605563640594, "train/lr": 1.486348372269683e-05, "perf/step_duration_sec": 3.479511406039819, "perf/samples_per_sec": 9.196693519800979, "perf/tokens_per_sec": 7550.7727764290985, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26273.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
178
+ {"timestamp": 1774541522.5523932, "event": "train_step", "step": 580, "epoch": 3, "metrics": {"train/step_loss": 0.829232782125473, "train/step_real_loss": 0.829232782125473, "train/lr": 1.4092610048061092e-05, "perf/step_duration_sec": 3.3206318439915776, "perf/samples_per_sec": 9.636720209709933, "perf/tokens_per_sec": 7991.852528915069, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26538.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
179
+ {"timestamp": 1774541539.0717015, "event": "eval_step", "step": 580, "epoch": 3, "metrics": {"eval/loss": 0.75683419600777, "eval/duration_sec": 16.504327062051743}}
180
+ {"timestamp": 1774541556.3249874, "event": "train_step", "step": 585, "epoch": 3, "metrics": {"train/step_loss": 0.7542973458766937, "train/step_real_loss": 0.7542973458766937, "train/lr": 1.3369820871537878e-05, "perf/step_duration_sec": 3.5233090959955007, "perf/samples_per_sec": 9.082370898531256, "perf/tokens_per_sec": 7383.967540505911, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26016.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
181
+ {"timestamp": 1774541573.4270687, "event": "train_step", "step": 590, "epoch": 3, "metrics": {"train/step_loss": 0.7202246487140656, "train/step_real_loss": 0.7202246487140656, "train/lr": 1.2695578622102097e-05, "perf/step_duration_sec": 3.3558571129105985, "perf/samples_per_sec": 9.535566897914135, "perf/tokens_per_sec": 8611.510868213143, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28899.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
182
+ {"timestamp": 1774541589.3482428, "event": "eval_step", "step": 590, "epoch": 3, "metrics": {"eval/loss": 0.756835603195688, "eval/duration_sec": 15.901251000119373}}
183
+ {"timestamp": 1774541606.1840398, "event": "train_step", "step": 595, "epoch": 3, "metrics": {"train/step_loss": 0.8010530471801758, "train/step_real_loss": 0.8010530471801758, "train/lr": 1.207031466917989e-05, "perf/step_duration_sec": 3.380698140943423, "perf/samples_per_sec": 9.465500516727598, "perf/tokens_per_sec": 7646.645433060161, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25851.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
184
+ {"timestamp": 1774541623.3525214, "event": "train_step", "step": 600, "epoch": 3, "metrics": {"train/step_loss": 0.782949686050415, "train/step_real_loss": 0.782949686050415, "train/lr": 1.1494429046665192e-05, "perf/step_duration_sec": 3.53925053589046, "perf/samples_per_sec": 9.041462217918104, "perf/tokens_per_sec": 7991.522417862365, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 28284.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
185
+ {"timestamp": 1774541640.6541808, "event": "eval_step", "step": 600, "epoch": 3, "metrics": {"eval/loss": 0.756805557757616, "eval/duration_sec": 17.283081082161516}}
186
+ {"timestamp": 1774541657.8988686, "event": "train_step", "step": 605, "epoch": 3, "metrics": {"train/step_loss": 0.7629248797893524, "train/step_real_loss": 0.7629248797893524, "train/lr": 1.09682901969846e-05, "perf/step_duration_sec": 3.429724073037505, "perf/samples_per_sec": 9.330196633474214, "perf/tokens_per_sec": 7525.678290831342, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25811.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
187
+ {"timestamp": 1774541674.9353623, "event": "train_step", "step": 610, "epoch": 3, "metrics": {"train/step_loss": 0.7820979952812195, "train/step_real_loss": 0.7820979952812195, "train/lr": 1.049223473537368e-05, "perf/step_duration_sec": 3.3810153561644256, "perf/samples_per_sec": 9.464612440063634, "perf/tokens_per_sec": 7073.023184115054, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 23914.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
188
+ {"timestamp": 1774541692.163433, "event": "eval_step", "step": 610, "epoch": 3, "metrics": {"eval/loss": 0.7566967834920986, "eval/duration_sec": 17.22026076889597}}
189
+ {"timestamp": 1774541709.0971062, "event": "train_step", "step": 615, "epoch": 3, "metrics": {"train/step_loss": 0.7408742904663086, "train/step_real_loss": 0.7408742904663086, "train/lr": 1.0066567234516006e-05, "perf/step_duration_sec": 3.499847089871764, "perf/samples_per_sec": 9.143256599011156, "perf/tokens_per_sec": 7680.335543169372, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26880.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
190
+ {"timestamp": 1774541726.1687822, "event": "train_step", "step": 620, "epoch": 3, "metrics": {"train/step_loss": 0.7260139584541321, "train/step_real_loss": 0.7260139584541321, "train/lr": 9.691560029682284e-06, "perf/step_duration_sec": 3.336881665047258, "perf/samples_per_sec": 9.589791671424706, "perf/tokens_per_sec": 8001.78210683472, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26701.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
191
+ {"timestamp": 1774541742.8133547, "event": "eval_step", "step": 620, "epoch": 3, "metrics": {"eval/loss": 0.7566954752673274, "eval/duration_sec": 16.6284677030053}}
192
+ {"timestamp": 1774541759.9906392, "event": "train_step", "step": 625, "epoch": 3, "metrics": {"train/step_loss": 0.790774941444397, "train/step_real_loss": 0.790774941444397, "train/lr": 9.367453044494745e-06, "perf/step_duration_sec": 3.373224897077307, "perf/samples_per_sec": 9.486470951796319, "perf/tokens_per_sec": 7261.597061382839, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 24495.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
193
+ {"timestamp": 1774541777.25411, "event": "train_step", "step": 630, "epoch": 3, "metrics": {"train/step_loss": 0.7545097768306732, "train/step_real_loss": 0.7545097768306732, "train/lr": 9.094453637427747e-06, "perf/step_duration_sec": 3.5466750538907945, "perf/samples_per_sec": 9.022535054316625, "perf/tokens_per_sec": 7208.159645737642, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25565.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39158344268799, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
194
+ {"timestamp": 1774541794.056787, "event": "eval_step", "step": 630, "epoch": 3, "metrics": {"eval/loss": 0.7566965343835561, "eval/duration_sec": 16.790947048924863}}
195
+ {"timestamp": 1774541811.431628, "event": "train_step", "step": 635, "epoch": 3, "metrics": {"train/step_loss": 0.7455289363861084, "train/step_real_loss": 0.7455289363861084, "train/lr": 8.872736469143175e-06, "perf/step_duration_sec": 3.3892036308534443, "perf/samples_per_sec": 9.441746051694745, "perf/tokens_per_sec": 7819.8311127520565, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 26503.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
196
+ {"timestamp": 1774541828.5450737, "event": "train_step", "step": 640, "epoch": 3, "metrics": {"train/step_loss": 0.7634424567222595, "train/step_real_loss": 0.7634424567222595, "train/lr": 8.7024433907453e-06, "perf/step_duration_sec": 3.395832329057157, "perf/samples_per_sec": 9.423315670265943, "perf/tokens_per_sec": 7992.444081458998, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 27141.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
197
+ {"timestamp": 1774541844.9945858, "event": "eval_step", "step": 640, "epoch": 3, "metrics": {"eval/loss": 0.7567026015856994, "eval/duration_sec": 16.43775229086168}}
198
+ {"timestamp": 1774541861.8458343, "event": "train_step", "step": 645, "epoch": 3, "metrics": {"train/step_loss": 0.726156622171402, "train/step_real_loss": 0.726156622171402, "train/lr": 8.583683353026752e-06, "perf/step_duration_sec": 3.4520575769711286, "perf/samples_per_sec": 9.269833798101692, "perf/tokens_per_sec": 7483.073333517591, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25832.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
199
+ {"timestamp": 1774541878.9150581, "event": "train_step", "step": 650, "epoch": 3, "metrics": {"train/step_loss": 0.7592037320137024, "train/step_real_loss": 0.7592037320137024, "train/lr": 8.516532336763516e-06, "perf/step_duration_sec": 3.4415706910658628, "perf/samples_per_sec": 9.298080112975835, "perf/tokens_per_sec": 7483.2111009243645, "perf/logical_batch_size": 32.0, "perf/logical_token_count": 25754.0, "perf/gradient_accumulation_steps": 2.0, "system/cuda_memory_allocated_gb": 18.39121723175049, "system/cuda_max_memory_allocated_gb": 93.7962327003479}}
200
+ {"timestamp": 1774541895.3553648, "event": "eval_step", "step": 650, "epoch": 3, "metrics": {"eval/loss": 0.7566873039240423, "eval/duration_sec": 16.427802158053964}}
201
+ {"timestamp": 1774541925.5017893, "event": "train_epoch", "step": 654, "epoch": 3, "metrics": {"train/epoch_loss": 0.7756987899507034, "train/epoch_real_loss": 0.7591156756009804, "train/epoch_canary_loss": 6.751691017316016, "perf/epoch_duration_sec": 1114.2803215740714, "perf/epoch_samples_per_sec": 50.227037951221334, "perf/epoch_tokens_per_sec": 42010.19087716901, "perf/epoch_samples": 55967.0, "perf/epoch_tokens": 46811129.0, "system/cuda_epoch_peak_memory_gb": 93.7962327003479, "eval/loss": 0.756622653292573, "eval/duration_sec": 15.878157561877742}}
202
+ {"timestamp": 1774541936.2744477, "event": "audit_epoch", "step": 654, "epoch": 3, "metrics": {"audit/delta": 1e-05, "audit/num_canaries": 500.0, "audit/num_members": 250.0, "audit/paper_guess_fraction": 0.2, "audit/paper_guess_steps": 20.0, "audit/loss/auc": 0.924608, "audit/loss/empirical_epsilon/0.05": 3.4791953936219215, "audit/loss/empirical_epsilon/0.01": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/loss/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 100.0, "audit/embedding/auc": 0.756, "audit/embedding/empirical_epsilon/0.05": 3.4791953936219215, "audit/embedding/empirical_epsilon/0.01": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 100.0, "perf/audit_duration_sec": 8.446189163951203}}
203
+ {"timestamp": 1774541946.0076194, "event": "audit_final", "step": 654, "epoch": 3, "metrics": {"audit/delta": 1e-05, "audit/num_canaries": 500.0, "audit/num_members": 250.0, "audit/paper_guess_fraction": 0.2, "audit/paper_guess_steps": 20.0, "audit/loss/auc": 0.922912, "audit/loss/empirical_epsilon/0.05": 3.4791953936219215, "audit/loss/empirical_epsilon/0.01": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/loss/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/loss/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 100.0, "audit/embedding/auc": 0.756, "audit/embedding/empirical_epsilon/0.05": 3.4791953936219215, "audit/embedding/empirical_epsilon/0.01": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215, "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/epsilon": 3.023197554051876, "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 100.0, "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 100.0}}
204
+ {"timestamp": 1774541946.569153, "event": "energy_final", "step": 654, "epoch": null, "metrics": {"energy/codecarbon/duration": 3649.28405967704, "energy/codecarbon/emissions": 0.09680917842542971, "energy/codecarbon/emissions_rate": 2.652826604953227e-05, "energy/codecarbon/cpu_power": 80.67234615823537, "energy/codecarbon/gpu_power": 2632.7542922341963, "energy/codecarbon/ram_power": 38.0, "energy/codecarbon/cpu_energy": 0.07865241538360482, "energy/codecarbon/gpu_energy": 2.6627394257454142, "energy/codecarbon/ram_energy": 0.037047972475709026, "energy/codecarbon/energy_consumed": 2.7784398136047317, "energy/codecarbon/water_consumed": 0.0, "energy/codecarbon/cpu_count": 8.0, "energy/codecarbon/gpu_count": 8.0, "energy/codecarbon/longitude": 8.212, "energy/codecarbon/latitude": 47.4843, "energy/codecarbon/ram_total_size": 256.0, "energy/codecarbon/cpu_utilization_percent": 11.109346567411084, "energy/codecarbon/gpu_utilization_percent": 69.25771987868762, "energy/codecarbon/ram_utilization_percent": 26.639178384339676, "energy/codecarbon/ram_used_gb": 528.0563867020483, "energy/codecarbon/pue": 1.0, "energy/codecarbon/wue": 0.0}}
granite-4.0-h-tiny/base_attn/resolved_config.yaml ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ name: ibm-granite/granite-4.0-h-tiny-base
3
+ tokenizer_name: ibm-granite/granite-4.0-h-tiny-base
4
+ max_length: 1024
5
+ dtype: bfloat16
6
+ trust_remote_code: true
7
+ use_fast_tokenizer: true
8
+ cache_dir: null
9
+ local_files_only: false
10
+ low_cpu_mem_usage: true
11
+ tie_word_embeddings: true
12
+ gradient_checkpointing: false
13
+ use_chat_template: false
14
+ dataset:
15
+ name: melihcatal/codedp-cpt
16
+ split: train
17
+ mode: cpt
18
+ text_column: text
19
+ validation_ratio: 0.05
20
+ max_samples: -1
21
+ lora:
22
+ enabled: true
23
+ r: 16
24
+ alpha: 32
25
+ dropout: 0.05
26
+ target_modules:
27
+ - q_proj
28
+ - k_proj
29
+ - v_proj
30
+ - o_proj
31
+ modules_to_save:
32
+ - lm_head
33
+ bias: none
34
+ training:
35
+ seed: 42
36
+ epochs: 3
37
+ warmup_steps: null
38
+ warmup_ratio: 0.05
39
+ mixed_precision: false
40
+ mixed_precision_dtype: bfloat16
41
+ batch_size: 16
42
+ eval_batch_size: 8
43
+ eval_every_steps: 10
44
+ eval_every_epochs: 1
45
+ learning_rate: 0.00017
46
+ optimizer: adamw
47
+ lr_scheduler: cosine
48
+ adam_beta1: 0.9
49
+ adam_beta2: 0.999
50
+ adam_epsilon: 1.0e-08
51
+ sgd_momentum: 0.9
52
+ weight_decay: 0.01
53
+ max_grad_norm: 1.0
54
+ log_every: 5
55
+ gradient_accumulation_steps: 2
56
+ num_workers: 4
57
+ output_dir: /scratch/mcatal/runs/cpt/granite-4.0-h-tiny/base_attn
58
+ min_lr_ratio: 0.05
59
+ distributed:
60
+ strategy: dpddp
61
+ backend: nccl
62
+ devices: null
63
+ dp:
64
+ module_validator: auto
65
+ target_delta: 1.0e-05
66
+ noise_multiplier: null
67
+ max_grad_norm: 1.0
68
+ grad_sample_mode: hooks
69
+ clipping: flat
70
+ secure_mode: false
71
+ enabled: false
72
+ target_epsilon: 8.0
73
+ audit:
74
+ enabled: true
75
+ run_every_epoch: true
76
+ epoch_device: cuda
77
+ q_canary: auto
78
+ num_canaries: 500
79
+ prefix_length: 49
80
+ num_digits: 12
81
+ batch_size: 32
82
+ delta: 1.0e-05
83
+ p_values:
84
+ - 0.05
85
+ - 0.01
86
+ paper_guess_fraction: 0.2
87
+ paper_guess_steps: 20
88
+ enable_holdout_empirical_epsilon: false
89
+ holdout_seed: 42
90
+ tie_seed: 42
91
+ tracking:
92
+ enabled: true
93
+ tensorboard: true
94
+ wandb: false
95
+ wandb_project: codedp-finetune-h200-audit
96
+ wandb_run_name: granite-4.0-h-tiny-cpt-base_attn
97
+ wandb_mode: online
98
+ codecarbon: true
99
+ codecarbon_output_file: codecarbon.csv
100
+ codecarbon_measure_power_secs: 15
101
+ codecarbon_country_iso_code: null
102
+ codecarbon_project_name: codedp-granite-4.0-h-tiny-cpt-base_attn
103
+ moe:
104
+ output_router_logits: false
105
+ router_aux_loss_coef: 0.0
106
+ freeze_router: true
107
+ profile:
108
+ enabled: false
109
+ num_batches: 8
110
+ top_experts: 8
111
+ output_file: moe_expert_profile.json
granite-4.0-h-tiny/base_attn/summary.json ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audit/delta": 1e-05,
3
+ "audit/embedding/auc": 0.756,
4
+ "audit/embedding/empirical_epsilon/0.01": 3.023197554051876,
5
+ "audit/embedding/empirical_epsilon/0.05": 3.4791953936219215,
6
+ "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 100.0,
7
+ "audit/embedding/empirical_epsilon_details/0.01/epsilon": 3.023197554051876,
8
+ "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 100.0,
9
+ "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 100.0,
10
+ "audit/embedding/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215,
11
+ "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 100.0,
12
+ "audit/loss/auc": 0.922912,
13
+ "audit/loss/empirical_epsilon/0.01": 3.023197554051876,
14
+ "audit/loss/empirical_epsilon/0.05": 3.4791953936219215,
15
+ "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 100.0,
16
+ "audit/loss/empirical_epsilon_details/0.01/epsilon": 3.023197554051876,
17
+ "audit/loss/empirical_epsilon_details/0.01/num_guesses": 100.0,
18
+ "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 100.0,
19
+ "audit/loss/empirical_epsilon_details/0.05/epsilon": 3.4791953936219215,
20
+ "audit/loss/empirical_epsilon_details/0.05/num_guesses": 100.0,
21
+ "audit/num_canaries": 500.0,
22
+ "audit/num_members": 250.0,
23
+ "audit/paper_guess_fraction": 0.2,
24
+ "audit/paper_guess_steps": 20.0,
25
+ "energy/codecarbon/cpu_count": 8.0,
26
+ "energy/codecarbon/cpu_energy": 0.07865241538360482,
27
+ "energy/codecarbon/cpu_power": 80.67234615823537,
28
+ "energy/codecarbon/cpu_utilization_percent": 11.109346567411084,
29
+ "energy/codecarbon/duration": 3649.28405967704,
30
+ "energy/codecarbon/emissions": 0.09680917842542971,
31
+ "energy/codecarbon/emissions_rate": 2.652826604953227e-05,
32
+ "energy/codecarbon/energy_consumed": 2.7784398136047317,
33
+ "energy/codecarbon/gpu_count": 8.0,
34
+ "energy/codecarbon/gpu_energy": 2.6627394257454142,
35
+ "energy/codecarbon/gpu_power": 2632.7542922341963,
36
+ "energy/codecarbon/gpu_utilization_percent": 69.25771987868762,
37
+ "energy/codecarbon/latitude": 47.4843,
38
+ "energy/codecarbon/longitude": 8.212,
39
+ "energy/codecarbon/pue": 1.0,
40
+ "energy/codecarbon/ram_energy": 0.037047972475709026,
41
+ "energy/codecarbon/ram_power": 38.0,
42
+ "energy/codecarbon/ram_total_size": 256.0,
43
+ "energy/codecarbon/ram_used_gb": 528.0563867020483,
44
+ "energy/codecarbon/ram_utilization_percent": 26.639178384339676,
45
+ "energy/codecarbon/water_consumed": 0.0,
46
+ "energy/codecarbon/wue": 0.0,
47
+ "eval/duration_sec": 15.878157561877742,
48
+ "eval/loss": 0.756622653292573,
49
+ "perf/audit_duration_sec": 8.446189163951203,
50
+ "perf/epoch_duration_sec": 1114.2803215740714,
51
+ "perf/epoch_samples": 55967.0,
52
+ "perf/epoch_samples_per_sec": 50.227037951221334,
53
+ "perf/epoch_tokens": 46811129.0,
54
+ "perf/epoch_tokens_per_sec": 42010.19087716901,
55
+ "perf/gradient_accumulation_steps": 2.0,
56
+ "perf/logical_batch_size": 32.0,
57
+ "perf/logical_token_count": 25754.0,
58
+ "perf/samples_per_sec": 9.298080112975835,
59
+ "perf/step_duration_sec": 3.4415706910658628,
60
+ "perf/tokens_per_sec": 7483.2111009243645,
61
+ "system/cuda_epoch_peak_memory_gb": 93.7962327003479,
62
+ "system/cuda_max_memory_allocated_gb": 93.7962327003479,
63
+ "system/cuda_memory_allocated_gb": 18.39121723175049,
64
+ "train/epoch_canary_loss": 6.751691017316016,
65
+ "train/epoch_loss": 0.7756987899507034,
66
+ "train/epoch_real_loss": 0.7591156756009804,
67
+ "train/lr": 8.516532336763516e-06,
68
+ "train/step_canary_loss": 4.03125,
69
+ "train/step_loss": 0.7592037320137024,
70
+ "train/step_real_loss": 0.7592037320137024
71
+ }
granite-4.0-h-tiny/base_attn/tokenizer/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/base_attn/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,516 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|end_of_text|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|end_of_text|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": {
9
+ "extra_special_token_0": "865331112869",
10
+ "extra_special_token_1": "569765693871",
11
+ "extra_special_token_2": "485177821815",
12
+ "extra_special_token_3": "135441121756",
13
+ "extra_special_token_4": "367459894796",
14
+ "extra_special_token_5": "877482678543",
15
+ "extra_special_token_6": "457919547633",
16
+ "extra_special_token_7": "765474393376",
17
+ "extra_special_token_8": "114848338811",
18
+ "extra_special_token_9": "746285987371",
19
+ "extra_special_token_10": "649291669397",
20
+ "extra_special_token_11": "927914615679",
21
+ "extra_special_token_12": "445925149649",
22
+ "extra_special_token_13": "691587454538",
23
+ "extra_special_token_14": "143777992227",
24
+ "extra_special_token_15": "997981281989",
25
+ "extra_special_token_16": "425949483533",
26
+ "extra_special_token_17": "982993456429",
27
+ "extra_special_token_18": "718726519731",
28
+ "extra_special_token_19": "172599315861",
29
+ "extra_special_token_20": "643489267333",
30
+ "extra_special_token_21": "282322838685",
31
+ "extra_special_token_22": "781653545886",
32
+ "extra_special_token_23": "796415361892",
33
+ "extra_special_token_24": "841991688488",
34
+ "extra_special_token_25": "211411365397",
35
+ "extra_special_token_26": "698218415444",
36
+ "extra_special_token_27": "355977139358",
37
+ "extra_special_token_28": "682564697312",
38
+ "extra_special_token_29": "383837596997",
39
+ "extra_special_token_30": "689362171782",
40
+ "extra_special_token_31": "749966767285",
41
+ "extra_special_token_32": "753159165157",
42
+ "extra_special_token_33": "795693824762",
43
+ "extra_special_token_34": "669689115557",
44
+ "extra_special_token_35": "327491773134",
45
+ "extra_special_token_36": "983569279932",
46
+ "extra_special_token_37": "612128769512",
47
+ "extra_special_token_38": "374327157578",
48
+ "extra_special_token_39": "311632789559",
49
+ "extra_special_token_40": "523918658846",
50
+ "extra_special_token_41": "765981581453",
51
+ "extra_special_token_42": "794825141891",
52
+ "extra_special_token_43": "873898736873",
53
+ "extra_special_token_44": "447445629421",
54
+ "extra_special_token_45": "473822473819",
55
+ "extra_special_token_46": "181439694557",
56
+ "extra_special_token_47": "592538279337",
57
+ "extra_special_token_48": "668134915514",
58
+ "extra_special_token_49": "643692393748",
59
+ "extra_special_token_50": "696651276628",
60
+ "extra_special_token_51": "853859348234",
61
+ "extra_special_token_52": "778466723723",
62
+ "extra_special_token_53": "929826356991",
63
+ "extra_special_token_54": "272362973463",
64
+ "extra_special_token_55": "694235616268",
65
+ "extra_special_token_56": "281673864127",
66
+ "extra_special_token_57": "479676316326",
67
+ "extra_special_token_58": "646979124677",
68
+ "extra_special_token_59": "922327493433",
69
+ "extra_special_token_60": "883685933161",
70
+ "extra_special_token_61": "264259917554",
71
+ "extra_special_token_62": "836746273134",
72
+ "extra_special_token_63": "658481324922",
73
+ "extra_special_token_64": "481884157827",
74
+ "extra_special_token_65": "587787496812",
75
+ "extra_special_token_66": "579184949249",
76
+ "extra_special_token_67": "912193598348",
77
+ "extra_special_token_68": "529679678956",
78
+ "extra_special_token_69": "795838284624",
79
+ "extra_special_token_70": "159337222655",
80
+ "extra_special_token_71": "173781362446",
81
+ "extra_special_token_72": "773687856563",
82
+ "extra_special_token_73": "535787224917",
83
+ "extra_special_token_74": "351885857332",
84
+ "extra_special_token_75": "578827344666",
85
+ "extra_special_token_76": "198462689911",
86
+ "extra_special_token_77": "722618266242",
87
+ "extra_special_token_78": "952872416512",
88
+ "extra_special_token_79": "517778845323",
89
+ "extra_special_token_80": "749665846687",
90
+ "extra_special_token_81": "661436365453",
91
+ "extra_special_token_82": "259666844669",
92
+ "extra_special_token_83": "242851284913",
93
+ "extra_special_token_84": "514532995959",
94
+ "extra_special_token_85": "161588262349",
95
+ "extra_special_token_86": "742765629356",
96
+ "extra_special_token_87": "225164373623",
97
+ "extra_special_token_88": "676539973863",
98
+ "extra_special_token_89": "826214551218",
99
+ "extra_special_token_90": "182345464792",
100
+ "extra_special_token_91": "232776999554",
101
+ "extra_special_token_92": "337326533813",
102
+ "extra_special_token_93": "676676697292",
103
+ "extra_special_token_94": "929185622831",
104
+ "extra_special_token_95": "545512344383",
105
+ "extra_special_token_96": "499444466686",
106
+ "extra_special_token_97": "314697386682",
107
+ "extra_special_token_98": "517379856925",
108
+ "extra_special_token_99": "379557332953",
109
+ "extra_special_token_100": "614797267726",
110
+ "extra_special_token_101": "429781429464",
111
+ "extra_special_token_102": "922466849763",
112
+ "extra_special_token_103": "721737645236",
113
+ "extra_special_token_104": "479227349997",
114
+ "extra_special_token_105": "136931728327",
115
+ "extra_special_token_106": "259533577263",
116
+ "extra_special_token_107": "488538864842",
117
+ "extra_special_token_108": "937495658852",
118
+ "extra_special_token_109": "489991411364",
119
+ "extra_special_token_110": "499148455254",
120
+ "extra_special_token_111": "441373944925",
121
+ "extra_special_token_112": "899151413682",
122
+ "extra_special_token_113": "467893531755",
123
+ "extra_special_token_114": "527117488925",
124
+ "extra_special_token_115": "928335588653",
125
+ "extra_special_token_116": "374439448821",
126
+ "extra_special_token_117": "879425227932",
127
+ "extra_special_token_118": "867678158885",
128
+ "extra_special_token_119": "399749397872",
129
+ "extra_special_token_120": "129693547287",
130
+ "extra_special_token_121": "689285841825",
131
+ "extra_special_token_122": "771619544974",
132
+ "extra_special_token_123": "724883568652",
133
+ "extra_special_token_124": "516968424863",
134
+ "extra_special_token_125": "733737988257",
135
+ "extra_special_token_126": "852347289392",
136
+ "extra_special_token_127": "296953381169",
137
+ "extra_special_token_128": "377273562477",
138
+ "extra_special_token_129": "262296912232",
139
+ "extra_special_token_130": "547149832394",
140
+ "extra_special_token_131": "298464134954",
141
+ "extra_special_token_132": "216667245274",
142
+ "extra_special_token_133": "843998562287",
143
+ "extra_special_token_134": "572154333646",
144
+ "extra_special_token_135": "124589118494",
145
+ "extra_special_token_136": "841824384614",
146
+ "extra_special_token_137": "232896526252",
147
+ "extra_special_token_138": "295448593321",
148
+ "extra_special_token_139": "123741461297",
149
+ "extra_special_token_140": "653573457168",
150
+ "extra_special_token_141": "196735786156",
151
+ "extra_special_token_142": "377338713663",
152
+ "extra_special_token_143": "964342468552",
153
+ "extra_special_token_144": "586855179568",
154
+ "extra_special_token_145": "484773717614",
155
+ "extra_special_token_146": "894885246797",
156
+ "extra_special_token_147": "677896358599",
157
+ "extra_special_token_148": "848845611563",
158
+ "extra_special_token_149": "851852651677",
159
+ "extra_special_token_150": "398549545767",
160
+ "extra_special_token_151": "454244839926",
161
+ "extra_special_token_152": "799364566435",
162
+ "extra_special_token_153": "967114116556",
163
+ "extra_special_token_154": "817378986438",
164
+ "extra_special_token_155": "233795848681",
165
+ "extra_special_token_156": "824387273757",
166
+ "extra_special_token_157": "916198946615",
167
+ "extra_special_token_158": "563117729724",
168
+ "extra_special_token_159": "951794811935",
169
+ "extra_special_token_160": "374598961236",
170
+ "extra_special_token_161": "922867396683",
171
+ "extra_special_token_162": "765737843639",
172
+ "extra_special_token_163": "175469284871",
173
+ "extra_special_token_164": "231853711778",
174
+ "extra_special_token_165": "662426712668",
175
+ "extra_special_token_166": "711412347158",
176
+ "extra_special_token_167": "753466987363",
177
+ "extra_special_token_168": "513361312532",
178
+ "extra_special_token_169": "712992815957",
179
+ "extra_special_token_170": "971621888444",
180
+ "extra_special_token_171": "829235161526",
181
+ "extra_special_token_172": "585544633356",
182
+ "extra_special_token_173": "582471228164",
183
+ "extra_special_token_174": "678666359123",
184
+ "extra_special_token_175": "557533689478",
185
+ "extra_special_token_176": "632962475133",
186
+ "extra_special_token_177": "484489193824",
187
+ "extra_special_token_178": "489562189822",
188
+ "extra_special_token_179": "589547936288",
189
+ "extra_special_token_180": "363214487524",
190
+ "extra_special_token_181": "244885399387",
191
+ "extra_special_token_182": "431751228368",
192
+ "extra_special_token_183": "433581868192",
193
+ "extra_special_token_184": "486391569221",
194
+ "extra_special_token_185": "185438575221",
195
+ "extra_special_token_186": "126574388585",
196
+ "extra_special_token_187": "741757479784",
197
+ "extra_special_token_188": "529854679937",
198
+ "extra_special_token_189": "996116119839",
199
+ "extra_special_token_190": "616248973917",
200
+ "extra_special_token_191": "763531783491",
201
+ "extra_special_token_192": "955456118295",
202
+ "extra_special_token_193": "364196983365",
203
+ "extra_special_token_194": "195792996468",
204
+ "extra_special_token_195": "151859598873",
205
+ "extra_special_token_196": "399223169721",
206
+ "extra_special_token_197": "938488813964",
207
+ "extra_special_token_198": "961981959227",
208
+ "extra_special_token_199": "183368827562",
209
+ "extra_special_token_200": "533417736566",
210
+ "extra_special_token_201": "786391632558",
211
+ "extra_special_token_202": "665661658354",
212
+ "extra_special_token_203": "693281533643",
213
+ "extra_special_token_204": "475794684356",
214
+ "extra_special_token_205": "652154162978",
215
+ "extra_special_token_206": "753233719644",
216
+ "extra_special_token_207": "668514843129",
217
+ "extra_special_token_208": "819162623892",
218
+ "extra_special_token_209": "941169431859",
219
+ "extra_special_token_210": "877385381798",
220
+ "extra_special_token_211": "752644929761",
221
+ "extra_special_token_212": "881136466196",
222
+ "extra_special_token_213": "275597777299",
223
+ "extra_special_token_214": "731681792655",
224
+ "extra_special_token_215": "961133895172",
225
+ "extra_special_token_216": "864718285734",
226
+ "extra_special_token_217": "963852916563",
227
+ "extra_special_token_218": "319584985416",
228
+ "extra_special_token_219": "563365646341",
229
+ "extra_special_token_220": "811371928234",
230
+ "extra_special_token_221": "837131396371",
231
+ "extra_special_token_222": "267514771964",
232
+ "extra_special_token_223": "944513428457",
233
+ "extra_special_token_224": "117298239631",
234
+ "extra_special_token_225": "158142752582",
235
+ "extra_special_token_226": "252867443568",
236
+ "extra_special_token_227": "839269684865",
237
+ "extra_special_token_228": "612788593128",
238
+ "extra_special_token_229": "145669731981",
239
+ "extra_special_token_230": "121557291859",
240
+ "extra_special_token_231": "245416776926",
241
+ "extra_special_token_232": "799417897197",
242
+ "extra_special_token_233": "997958836435",
243
+ "extra_special_token_234": "892336777248",
244
+ "extra_special_token_235": "158929292238",
245
+ "extra_special_token_236": "581976444672",
246
+ "extra_special_token_237": "897784492783",
247
+ "extra_special_token_238": "492373714791",
248
+ "extra_special_token_239": "512659818733",
249
+ "extra_special_token_240": "881112998642",
250
+ "extra_special_token_241": "619454958782",
251
+ "extra_special_token_242": "431149748713",
252
+ "extra_special_token_243": "624221476921",
253
+ "extra_special_token_244": "125866399464",
254
+ "extra_special_token_245": "339882449689",
255
+ "extra_special_token_246": "186198784585",
256
+ "extra_special_token_247": "943193294691",
257
+ "extra_special_token_248": "955668961269",
258
+ "extra_special_token_249": "232787996724",
259
+ "extra_special_token_250": "215671314196",
260
+ "extra_special_token_251": "286173241916",
261
+ "extra_special_token_252": "745977673725",
262
+ "extra_special_token_253": "556976448182",
263
+ "extra_special_token_254": "599961512792",
264
+ "extra_special_token_255": "766294538337",
265
+ "extra_special_token_256": "934912591213",
266
+ "extra_special_token_257": "295118729589",
267
+ "extra_special_token_258": "529455466433",
268
+ "extra_special_token_259": "196119929397",
269
+ "extra_special_token_260": "379571934299",
270
+ "extra_special_token_261": "251789649997",
271
+ "extra_special_token_262": "564544131355",
272
+ "extra_special_token_263": "244371196654",
273
+ "extra_special_token_264": "384598329253",
274
+ "extra_special_token_265": "887753195844",
275
+ "extra_special_token_266": "364947325679",
276
+ "extra_special_token_267": "655517954651",
277
+ "extra_special_token_268": "673948786567",
278
+ "extra_special_token_269": "857231548835",
279
+ "extra_special_token_270": "816115936673",
280
+ "extra_special_token_271": "644234165531",
281
+ "extra_special_token_272": "182782912224",
282
+ "extra_special_token_273": "234316622259",
283
+ "extra_special_token_274": "421369185549",
284
+ "extra_special_token_275": "434632855397",
285
+ "extra_special_token_276": "921889371893",
286
+ "extra_special_token_277": "415956914763",
287
+ "extra_special_token_278": "598916996413",
288
+ "extra_special_token_279": "773671349113",
289
+ "extra_special_token_280": "952465217972",
290
+ "extra_special_token_281": "117657531962",
291
+ "extra_special_token_282": "729825168745",
292
+ "extra_special_token_283": "691315125346",
293
+ "extra_special_token_284": "768461952319",
294
+ "extra_special_token_285": "664847713559",
295
+ "extra_special_token_286": "953267689786",
296
+ "extra_special_token_287": "886464195129",
297
+ "extra_special_token_288": "824488329416",
298
+ "extra_special_token_289": "837873762491",
299
+ "extra_special_token_290": "532833541879",
300
+ "extra_special_token_291": "669183782449",
301
+ "extra_special_token_292": "941976537588",
302
+ "extra_special_token_293": "739394546916",
303
+ "extra_special_token_294": "267954879268",
304
+ "extra_special_token_295": "637551427887",
305
+ "extra_special_token_296": "217756494954",
306
+ "extra_special_token_297": "524444658383",
307
+ "extra_special_token_298": "117783274348",
308
+ "extra_special_token_299": "138218735276",
309
+ "extra_special_token_300": "814611949491",
310
+ "extra_special_token_301": "711641973413",
311
+ "extra_special_token_302": "499156317423",
312
+ "extra_special_token_303": "515856611931",
313
+ "extra_special_token_304": "454164859837",
314
+ "extra_special_token_305": "345271433112",
315
+ "extra_special_token_306": "462294118988",
316
+ "extra_special_token_307": "511785788222",
317
+ "extra_special_token_308": "497294727353",
318
+ "extra_special_token_309": "866519986723",
319
+ "extra_special_token_310": "334513529294",
320
+ "extra_special_token_311": "549946382131",
321
+ "extra_special_token_312": "284445431422",
322
+ "extra_special_token_313": "396521188476",
323
+ "extra_special_token_314": "421435255895",
324
+ "extra_special_token_315": "133373659361",
325
+ "extra_special_token_316": "322683334381",
326
+ "extra_special_token_317": "228358422847",
327
+ "extra_special_token_318": "291762694874",
328
+ "extra_special_token_319": "143182978129",
329
+ "extra_special_token_320": "511923256573",
330
+ "extra_special_token_321": "327158398268",
331
+ "extra_special_token_322": "879764613759",
332
+ "extra_special_token_323": "564395222747",
333
+ "extra_special_token_324": "451161679736",
334
+ "extra_special_token_325": "538631466654",
335
+ "extra_special_token_326": "221762325616",
336
+ "extra_special_token_327": "218391991184",
337
+ "extra_special_token_328": "322589379462",
338
+ "extra_special_token_329": "876537814263",
339
+ "extra_special_token_330": "152676556624",
340
+ "extra_special_token_331": "332522971941",
341
+ "extra_special_token_332": "884354318946",
342
+ "extra_special_token_333": "513349618943",
343
+ "extra_special_token_334": "116639746413",
344
+ "extra_special_token_335": "635185846287",
345
+ "extra_special_token_336": "993832498489",
346
+ "extra_special_token_337": "813981174797",
347
+ "extra_special_token_338": "438745114173",
348
+ "extra_special_token_339": "983493951323",
349
+ "extra_special_token_340": "724492262421",
350
+ "extra_special_token_341": "622553389126",
351
+ "extra_special_token_342": "889965243135",
352
+ "extra_special_token_343": "364492359246",
353
+ "extra_special_token_344": "154962668224",
354
+ "extra_special_token_345": "179564995814",
355
+ "extra_special_token_346": "418412875665",
356
+ "extra_special_token_347": "718951851413",
357
+ "extra_special_token_348": "699446724178",
358
+ "extra_special_token_349": "624266421831",
359
+ "extra_special_token_350": "815458725125",
360
+ "extra_special_token_351": "455423278865",
361
+ "extra_special_token_352": "393741199486",
362
+ "extra_special_token_353": "328552864359",
363
+ "extra_special_token_354": "211662639865",
364
+ "extra_special_token_355": "218784516525",
365
+ "extra_special_token_356": "762486672996",
366
+ "extra_special_token_357": "142799718159",
367
+ "extra_special_token_358": "858146415154",
368
+ "extra_special_token_359": "767858144912",
369
+ "extra_special_token_360": "571317457151",
370
+ "extra_special_token_361": "635127952696",
371
+ "extra_special_token_362": "116427191984",
372
+ "extra_special_token_363": "268921994538",
373
+ "extra_special_token_364": "523937669294",
374
+ "extra_special_token_365": "165429152138",
375
+ "extra_special_token_366": "739246183345",
376
+ "extra_special_token_367": "591464355756",
377
+ "extra_special_token_368": "212985874612",
378
+ "extra_special_token_369": "191887635211",
379
+ "extra_special_token_370": "967214577653",
380
+ "extra_special_token_371": "119342152414",
381
+ "extra_special_token_372": "946444632795",
382
+ "extra_special_token_373": "618423867817",
383
+ "extra_special_token_374": "228565148417",
384
+ "extra_special_token_375": "729116422489",
385
+ "extra_special_token_376": "527874729936",
386
+ "extra_special_token_377": "739784153482",
387
+ "extra_special_token_378": "387763951128",
388
+ "extra_special_token_379": "331369926711",
389
+ "extra_special_token_380": "562716493614",
390
+ "extra_special_token_381": "739667844957",
391
+ "extra_special_token_382": "562389434565",
392
+ "extra_special_token_383": "256497188281",
393
+ "extra_special_token_384": "859927364588",
394
+ "extra_special_token_385": "417668946583",
395
+ "extra_special_token_386": "357621613582",
396
+ "extra_special_token_387": "438435178228",
397
+ "extra_special_token_388": "485692541169",
398
+ "extra_special_token_389": "825815739116",
399
+ "extra_special_token_390": "342221452223",
400
+ "extra_special_token_391": "697747991249",
401
+ "extra_special_token_392": "716763689965",
402
+ "extra_special_token_393": "141499982867",
403
+ "extra_special_token_394": "818479319499",
404
+ "extra_special_token_395": "336813343298",
405
+ "extra_special_token_396": "594688742928",
406
+ "extra_special_token_397": "472129283475",
407
+ "extra_special_token_398": "514354144759",
408
+ "extra_special_token_399": "349249721685",
409
+ "extra_special_token_400": "546276298359",
410
+ "extra_special_token_401": "353755529131",
411
+ "extra_special_token_402": "315534574435",
412
+ "extra_special_token_403": "523723475786",
413
+ "extra_special_token_404": "215826764872",
414
+ "extra_special_token_405": "367968398551",
415
+ "extra_special_token_406": "569853653352",
416
+ "extra_special_token_407": "389715484387",
417
+ "extra_special_token_408": "293847485454",
418
+ "extra_special_token_409": "714738141818",
419
+ "extra_special_token_410": "178478368922",
420
+ "extra_special_token_411": "581493616981",
421
+ "extra_special_token_412": "589439538674",
422
+ "extra_special_token_413": "846657726193",
423
+ "extra_special_token_414": "722339992679",
424
+ "extra_special_token_415": "138154781148",
425
+ "extra_special_token_416": "757785319772",
426
+ "extra_special_token_417": "492516914298",
427
+ "extra_special_token_418": "919181521716",
428
+ "extra_special_token_419": "985781138935",
429
+ "extra_special_token_420": "476969195485",
430
+ "extra_special_token_421": "313145133463",
431
+ "extra_special_token_422": "758963111966",
432
+ "extra_special_token_423": "147541537162",
433
+ "extra_special_token_424": "557163366873",
434
+ "extra_special_token_425": "144373897488",
435
+ "extra_special_token_426": "522515164754",
436
+ "extra_special_token_427": "724964923582",
437
+ "extra_special_token_428": "284776712475",
438
+ "extra_special_token_429": "375429755114",
439
+ "extra_special_token_430": "181233596124",
440
+ "extra_special_token_431": "948585673431",
441
+ "extra_special_token_432": "243165586174",
442
+ "extra_special_token_433": "396847976144",
443
+ "extra_special_token_434": "997724962668",
444
+ "extra_special_token_435": "558837194455",
445
+ "extra_special_token_436": "163165456396",
446
+ "extra_special_token_437": "378749551722",
447
+ "extra_special_token_438": "161238482259",
448
+ "extra_special_token_439": "754978243758",
449
+ "extra_special_token_440": "195388849133",
450
+ "extra_special_token_441": "229775525672",
451
+ "extra_special_token_442": "262437452884",
452
+ "extra_special_token_443": "441377892146",
453
+ "extra_special_token_444": "451885565366",
454
+ "extra_special_token_445": "981277526855",
455
+ "extra_special_token_446": "762495822823",
456
+ "extra_special_token_447": "368763327262",
457
+ "extra_special_token_448": "757422791351",
458
+ "extra_special_token_449": "636324136426",
459
+ "extra_special_token_450": "214193645583",
460
+ "extra_special_token_451": "412843856172",
461
+ "extra_special_token_452": "179386156569",
462
+ "extra_special_token_453": "756916173536",
463
+ "extra_special_token_454": "892697125149",
464
+ "extra_special_token_455": "625334487352",
465
+ "extra_special_token_456": "941861857715",
466
+ "extra_special_token_457": "887417525236",
467
+ "extra_special_token_458": "649516938598",
468
+ "extra_special_token_459": "717628619782",
469
+ "extra_special_token_460": "438124184139",
470
+ "extra_special_token_461": "547563892268",
471
+ "extra_special_token_462": "856317483891",
472
+ "extra_special_token_463": "313313831273",
473
+ "extra_special_token_464": "371496153876",
474
+ "extra_special_token_465": "587541149322",
475
+ "extra_special_token_466": "265847332563",
476
+ "extra_special_token_467": "449549215429",
477
+ "extra_special_token_468": "163497196769",
478
+ "extra_special_token_469": "861342291298",
479
+ "extra_special_token_470": "268433315926",
480
+ "extra_special_token_471": "774679513717",
481
+ "extra_special_token_472": "851254219729",
482
+ "extra_special_token_473": "583527834464",
483
+ "extra_special_token_474": "488496781997",
484
+ "extra_special_token_475": "556814553861",
485
+ "extra_special_token_476": "482829231639",
486
+ "extra_special_token_477": "618878266619",
487
+ "extra_special_token_478": "147444452794",
488
+ "extra_special_token_479": "949235426629",
489
+ "extra_special_token_480": "357299947518",
490
+ "extra_special_token_481": "175528632226",
491
+ "extra_special_token_482": "645527857972",
492
+ "extra_special_token_483": "186872457894",
493
+ "extra_special_token_484": "552738847828",
494
+ "extra_special_token_485": "626748382482",
495
+ "extra_special_token_486": "921894985642",
496
+ "extra_special_token_487": "943878645871",
497
+ "extra_special_token_488": "859289776479",
498
+ "extra_special_token_489": "614583493135",
499
+ "extra_special_token_490": "933775286797",
500
+ "extra_special_token_491": "332234613346",
501
+ "extra_special_token_492": "325196781219",
502
+ "extra_special_token_493": "142526557681",
503
+ "extra_special_token_494": "356722692178",
504
+ "extra_special_token_495": "449318681694",
505
+ "extra_special_token_496": "687284547244",
506
+ "extra_special_token_497": "947262995132",
507
+ "extra_special_token_498": "893974619684",
508
+ "extra_special_token_499": "797238311233"
509
+ },
510
+ "is_local": false,
511
+ "model_max_length": 1000000000000000019884624838656,
512
+ "pad_token": "<|pad|>",
513
+ "padding_side": "left",
514
+ "tokenizer_class": "GPT2Tokenizer",
515
+ "unk_token": "<|unk|>"
516
+ }
granite-4.0-h-tiny/base_attn/train.log ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-03-26 16:21:45,150 [INFO] new_opacus_codex.train_steps: epoch=1 step=5 loss=3.3723
2
+ 2026-03-26 16:22:08,246 [INFO] new_opacus_codex.train_steps: epoch=1 step=10 loss=3.3326
3
+ 2026-03-26 16:22:24,344 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=10 eval_loss=3.1472 duration_sec=16.09
4
+ 2026-03-26 16:22:43,741 [INFO] new_opacus_codex.train_steps: epoch=1 step=15 loss=3.2541
5
+ 2026-03-26 16:23:02,323 [INFO] new_opacus_codex.train_steps: epoch=1 step=20 loss=2.9884
6
+ 2026-03-26 16:23:19,195 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=20 eval_loss=2.0249 duration_sec=16.86
7
+ 2026-03-26 16:23:37,333 [INFO] new_opacus_codex.train_steps: epoch=1 step=25 loss=2.2884
8
+ 2026-03-26 16:23:55,413 [INFO] new_opacus_codex.train_steps: epoch=1 step=30 loss=1.6011
9
+ 2026-03-26 16:24:12,087 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=30 eval_loss=1.1229 duration_sec=16.66
10
+ 2026-03-26 16:24:30,992 [INFO] new_opacus_codex.train_steps: epoch=1 step=35 loss=1.2635
11
+ 2026-03-26 16:24:48,904 [INFO] new_opacus_codex.train_steps: epoch=1 step=40 loss=1.0702
12
+ 2026-03-26 16:25:05,398 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=40 eval_loss=0.9160 duration_sec=16.49
13
+ 2026-03-26 16:25:22,931 [INFO] new_opacus_codex.train_steps: epoch=1 step=45 loss=1.0026
14
+ 2026-03-26 16:25:40,254 [INFO] new_opacus_codex.train_steps: epoch=1 step=50 loss=0.9995
15
+ 2026-03-26 16:25:56,799 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=50 eval_loss=0.8713 duration_sec=16.53
16
+ 2026-03-26 16:26:14,256 [INFO] new_opacus_codex.train_steps: epoch=1 step=55 loss=1.0128
17
+ 2026-03-26 16:26:31,484 [INFO] new_opacus_codex.train_steps: epoch=1 step=60 loss=0.9683
18
+ 2026-03-26 16:26:48,444 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=60 eval_loss=0.8477 duration_sec=16.94
19
+ 2026-03-26 16:27:06,064 [INFO] new_opacus_codex.train_steps: epoch=1 step=65 loss=0.9178
20
+ 2026-03-26 16:27:23,187 [INFO] new_opacus_codex.train_steps: epoch=1 step=70 loss=0.8877
21
+ 2026-03-26 16:27:39,553 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=70 eval_loss=0.8339 duration_sec=16.36
22
+ 2026-03-26 16:27:57,080 [INFO] new_opacus_codex.train_steps: epoch=1 step=75 loss=0.9071
23
+ 2026-03-26 16:28:14,402 [INFO] new_opacus_codex.train_steps: epoch=1 step=80 loss=0.9515
24
+ 2026-03-26 16:28:31,136 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=80 eval_loss=0.8256 duration_sec=16.72
25
+ 2026-03-26 16:28:48,715 [INFO] new_opacus_codex.train_steps: epoch=1 step=85 loss=0.9050
26
+ 2026-03-26 16:29:06,143 [INFO] new_opacus_codex.train_steps: epoch=1 step=90 loss=0.9153
27
+ 2026-03-26 16:29:22,886 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=90 eval_loss=0.8181 duration_sec=16.73
28
+ 2026-03-26 16:29:40,506 [INFO] new_opacus_codex.train_steps: epoch=1 step=95 loss=0.9292
29
+ 2026-03-26 16:29:58,440 [INFO] new_opacus_codex.train_steps: epoch=1 step=100 loss=0.8889
30
+ 2026-03-26 16:30:15,181 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=100 eval_loss=0.8118 duration_sec=16.72
31
+ 2026-03-26 16:30:32,851 [INFO] new_opacus_codex.train_steps: epoch=1 step=105 loss=0.8651
32
+ 2026-03-26 16:30:50,040 [INFO] new_opacus_codex.train_steps: epoch=1 step=110 loss=0.9300
33
+ 2026-03-26 16:31:06,994 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=110 eval_loss=0.8064 duration_sec=16.93
34
+ 2026-03-26 16:31:24,408 [INFO] new_opacus_codex.train_steps: epoch=1 step=115 loss=0.9219
35
+ 2026-03-26 16:31:41,925 [INFO] new_opacus_codex.train_steps: epoch=1 step=120 loss=0.8523
36
+ 2026-03-26 16:31:59,277 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=120 eval_loss=0.8019 duration_sec=17.34
37
+ 2026-03-26 16:32:16,534 [INFO] new_opacus_codex.train_steps: epoch=1 step=125 loss=0.8554
38
+ 2026-03-26 16:32:34,332 [INFO] new_opacus_codex.train_steps: epoch=1 step=130 loss=0.8586
39
+ 2026-03-26 16:32:51,858 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=130 eval_loss=0.7980 duration_sec=17.51
40
+ 2026-03-26 16:33:09,292 [INFO] new_opacus_codex.train_steps: epoch=1 step=135 loss=0.8460
41
+ 2026-03-26 16:33:27,175 [INFO] new_opacus_codex.train_steps: epoch=1 step=140 loss=0.8273
42
+ 2026-03-26 16:33:43,435 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=140 eval_loss=0.7951 duration_sec=16.25
43
+ 2026-03-26 16:34:01,086 [INFO] new_opacus_codex.train_steps: epoch=1 step=145 loss=0.8238
44
+ 2026-03-26 16:34:18,442 [INFO] new_opacus_codex.train_steps: epoch=1 step=150 loss=0.8105
45
+ 2026-03-26 16:34:34,641 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=150 eval_loss=0.7926 duration_sec=16.19
46
+ 2026-03-26 16:34:51,964 [INFO] new_opacus_codex.train_steps: epoch=1 step=155 loss=0.8484
47
+ 2026-03-26 16:35:09,217 [INFO] new_opacus_codex.train_steps: epoch=1 step=160 loss=0.9103
48
+ 2026-03-26 16:35:26,037 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=160 eval_loss=0.7896 duration_sec=16.81
49
+ 2026-03-26 16:35:43,276 [INFO] new_opacus_codex.train_steps: epoch=1 step=165 loss=0.8725
50
+ 2026-03-26 16:36:01,289 [INFO] new_opacus_codex.train_steps: epoch=1 step=170 loss=0.8189
51
+ 2026-03-26 16:36:17,425 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=170 eval_loss=0.7866 duration_sec=16.12
52
+ 2026-03-26 16:36:34,939 [INFO] new_opacus_codex.train_steps: epoch=1 step=175 loss=0.8190
53
+ 2026-03-26 16:36:52,500 [INFO] new_opacus_codex.train_steps: epoch=1 step=180 loss=0.8298
54
+ 2026-03-26 16:37:08,998 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=180 eval_loss=0.7842 duration_sec=16.48
55
+ 2026-03-26 16:37:26,471 [INFO] new_opacus_codex.train_steps: epoch=1 step=185 loss=0.8305
56
+ 2026-03-26 16:37:43,784 [INFO] new_opacus_codex.train_steps: epoch=1 step=190 loss=0.8097
57
+ 2026-03-26 16:38:00,834 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=190 eval_loss=0.7818 duration_sec=17.03
58
+ 2026-03-26 16:38:18,268 [INFO] new_opacus_codex.train_steps: epoch=1 step=195 loss=0.8076
59
+ 2026-03-26 16:38:35,684 [INFO] new_opacus_codex.train_steps: epoch=1 step=200 loss=0.8425
60
+ 2026-03-26 16:38:52,554 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=200 eval_loss=0.7796 duration_sec=16.86
61
+ 2026-03-26 16:39:09,637 [INFO] new_opacus_codex.train_steps: epoch=1 step=205 loss=0.8392
62
+ 2026-03-26 16:39:26,589 [INFO] new_opacus_codex.train_steps: epoch=1 step=210 loss=0.8356
63
+ 2026-03-26 16:39:42,815 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=210 eval_loss=0.7775 duration_sec=16.21
64
+ 2026-03-26 16:40:00,275 [INFO] new_opacus_codex.train_steps: epoch=1 step=215 loss=0.8360
65
+ 2026-03-26 16:40:45,378 [INFO] new_opacus_codex.train_steps: epoch=2 step=220 loss=0.8543
66
+ 2026-03-26 16:41:01,529 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=220 eval_loss=0.7759 duration_sec=16.14
67
+ 2026-03-26 16:41:18,311 [INFO] new_opacus_codex.train_steps: epoch=2 step=225 loss=0.8017
68
+ 2026-03-26 16:41:35,119 [INFO] new_opacus_codex.train_steps: epoch=2 step=230 loss=0.8281
69
+ 2026-03-26 16:41:51,812 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=230 eval_loss=0.7742 duration_sec=16.68
70
+ 2026-03-26 16:42:09,571 [INFO] new_opacus_codex.train_steps: epoch=2 step=235 loss=0.8282
71
+ 2026-03-26 16:42:27,139 [INFO] new_opacus_codex.train_steps: epoch=2 step=240 loss=0.8018
72
+ 2026-03-26 16:42:43,968 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=240 eval_loss=0.7729 duration_sec=16.81
73
+ 2026-03-26 16:43:01,289 [INFO] new_opacus_codex.train_steps: epoch=2 step=245 loss=0.7983
74
+ 2026-03-26 16:43:19,131 [INFO] new_opacus_codex.train_steps: epoch=2 step=250 loss=0.8249
75
+ 2026-03-26 16:43:35,038 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=250 eval_loss=0.7716 duration_sec=15.89
76
+ 2026-03-26 16:43:52,651 [INFO] new_opacus_codex.train_steps: epoch=2 step=255 loss=0.8857
77
+ 2026-03-26 16:44:09,396 [INFO] new_opacus_codex.train_steps: epoch=2 step=260 loss=0.8301
78
+ 2026-03-26 16:44:26,002 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=260 eval_loss=0.7709 duration_sec=16.59
79
+ 2026-03-26 16:44:44,119 [INFO] new_opacus_codex.train_steps: epoch=2 step=265 loss=0.7971
80
+ 2026-03-26 16:45:04,103 [INFO] new_opacus_codex.train_steps: epoch=2 step=270 loss=0.8074
81
+ 2026-03-26 16:45:21,215 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=270 eval_loss=0.7696 duration_sec=17.10
82
+ 2026-03-26 16:45:39,995 [INFO] new_opacus_codex.train_steps: epoch=2 step=275 loss=0.8225
83
+ 2026-03-26 16:45:59,649 [INFO] new_opacus_codex.train_steps: epoch=2 step=280 loss=0.8141
84
+ 2026-03-26 16:46:15,793 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=280 eval_loss=0.7682 duration_sec=16.13
85
+ 2026-03-26 16:46:34,195 [INFO] new_opacus_codex.train_steps: epoch=2 step=285 loss=0.8076
86
+ 2026-03-26 16:46:51,894 [INFO] new_opacus_codex.train_steps: epoch=2 step=290 loss=0.8353
87
+ 2026-03-26 16:47:08,350 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=290 eval_loss=0.7673 duration_sec=16.42
88
+ 2026-03-26 16:47:25,680 [INFO] new_opacus_codex.train_steps: epoch=2 step=295 loss=0.7736
89
+ 2026-03-26 16:47:42,690 [INFO] new_opacus_codex.train_steps: epoch=2 step=300 loss=0.7991
90
+ 2026-03-26 16:47:59,109 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=300 eval_loss=0.7664 duration_sec=16.41
91
+ 2026-03-26 16:48:16,513 [INFO] new_opacus_codex.train_steps: epoch=2 step=305 loss=0.8491
92
+ 2026-03-26 16:48:33,607 [INFO] new_opacus_codex.train_steps: epoch=2 step=310 loss=0.7955
93
+ 2026-03-26 16:48:49,948 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=310 eval_loss=0.7657 duration_sec=16.33
94
+ 2026-03-26 16:49:07,013 [INFO] new_opacus_codex.train_steps: epoch=2 step=315 loss=0.7609
95
+ 2026-03-26 16:49:24,380 [INFO] new_opacus_codex.train_steps: epoch=2 step=320 loss=0.7725
96
+ 2026-03-26 16:49:40,891 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=320 eval_loss=0.7646 duration_sec=16.49
97
+ 2026-03-26 16:49:58,700 [INFO] new_opacus_codex.train_steps: epoch=2 step=325 loss=0.8200
98
+ 2026-03-26 16:50:15,848 [INFO] new_opacus_codex.train_steps: epoch=2 step=330 loss=0.8261
99
+ 2026-03-26 16:50:32,586 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=330 eval_loss=0.7637 duration_sec=16.72
100
+ 2026-03-26 16:50:49,766 [INFO] new_opacus_codex.train_steps: epoch=2 step=335 loss=0.7973
101
+ 2026-03-26 16:51:06,697 [INFO] new_opacus_codex.train_steps: epoch=2 step=340 loss=0.7634
102
+ 2026-03-26 16:51:23,766 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=340 eval_loss=0.7628 duration_sec=17.05
103
+ 2026-03-26 16:51:40,633 [INFO] new_opacus_codex.train_steps: epoch=2 step=345 loss=0.7894
104
+ 2026-03-26 16:51:58,266 [INFO] new_opacus_codex.train_steps: epoch=2 step=350 loss=0.8571
105
+ 2026-03-26 16:52:14,801 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=350 eval_loss=0.7622 duration_sec=16.52
106
+ 2026-03-26 16:52:32,521 [INFO] new_opacus_codex.train_steps: epoch=2 step=355 loss=0.8189
107
+ 2026-03-26 16:52:49,947 [INFO] new_opacus_codex.train_steps: epoch=2 step=360 loss=0.8021
108
+ 2026-03-26 16:53:06,025 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=360 eval_loss=0.7614 duration_sec=16.06
109
+ 2026-03-26 16:53:23,700 [INFO] new_opacus_codex.train_steps: epoch=2 step=365 loss=0.8258
110
+ 2026-03-26 16:53:41,417 [INFO] new_opacus_codex.train_steps: epoch=2 step=370 loss=0.8163
111
+ 2026-03-26 16:53:58,255 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=370 eval_loss=0.7608 duration_sec=16.82
112
+ 2026-03-26 16:54:16,153 [INFO] new_opacus_codex.train_steps: epoch=2 step=375 loss=0.7960
113
+ 2026-03-26 16:54:33,867 [INFO] new_opacus_codex.train_steps: epoch=2 step=380 loss=0.7744
114
+ 2026-03-26 16:54:50,672 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=380 eval_loss=0.7602 duration_sec=16.79
115
+ 2026-03-26 16:55:07,955 [INFO] new_opacus_codex.train_steps: epoch=2 step=385 loss=0.7762
116
+ 2026-03-26 16:55:25,421 [INFO] new_opacus_codex.train_steps: epoch=2 step=390 loss=0.7653
117
+ 2026-03-26 16:55:42,435 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=390 eval_loss=0.7598 duration_sec=16.99
118
+ 2026-03-26 16:55:59,645 [INFO] new_opacus_codex.train_steps: epoch=2 step=395 loss=0.7757
119
+ 2026-03-26 16:56:17,132 [INFO] new_opacus_codex.train_steps: epoch=2 step=400 loss=0.8156
120
+ 2026-03-26 16:56:33,930 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=400 eval_loss=0.7594 duration_sec=16.78
121
+ 2026-03-26 16:56:51,410 [INFO] new_opacus_codex.train_steps: epoch=2 step=405 loss=0.7840
122
+ 2026-03-26 16:57:08,384 [INFO] new_opacus_codex.train_steps: epoch=2 step=410 loss=0.7737
123
+ 2026-03-26 16:57:24,902 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=410 eval_loss=0.7589 duration_sec=16.50
124
+ 2026-03-26 16:57:42,157 [INFO] new_opacus_codex.train_steps: epoch=2 step=415 loss=0.8006
125
+ 2026-03-26 16:57:59,764 [INFO] new_opacus_codex.train_steps: epoch=2 step=420 loss=0.7899
126
+ 2026-03-26 16:58:16,469 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=420 eval_loss=0.7586 duration_sec=16.69
127
+ 2026-03-26 16:58:33,345 [INFO] new_opacus_codex.train_steps: epoch=2 step=425 loss=0.8201
128
+ 2026-03-26 16:58:50,586 [INFO] new_opacus_codex.train_steps: epoch=2 step=430 loss=0.8153
129
+ 2026-03-26 16:59:06,728 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=430 eval_loss=0.7584 duration_sec=16.13
130
+ 2026-03-26 16:59:24,178 [INFO] new_opacus_codex.train_steps: epoch=2 step=435 loss=0.7862
131
+ 2026-03-26 17:00:09,485 [INFO] new_opacus_codex.train_steps: epoch=3 step=440 loss=0.7370
132
+ 2026-03-26 17:00:26,533 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=440 eval_loss=0.7580 duration_sec=17.04
133
+ 2026-03-26 17:00:43,624 [INFO] new_opacus_codex.train_steps: epoch=3 step=445 loss=0.7732
134
+ 2026-03-26 17:01:01,181 [INFO] new_opacus_codex.train_steps: epoch=3 step=450 loss=0.7729
135
+ 2026-03-26 17:01:17,628 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=450 eval_loss=0.7579 duration_sec=16.44
136
+ 2026-03-26 17:01:35,177 [INFO] new_opacus_codex.train_steps: epoch=3 step=455 loss=0.7872
137
+ 2026-03-26 17:01:52,623 [INFO] new_opacus_codex.train_steps: epoch=3 step=460 loss=0.8024
138
+ 2026-03-26 17:02:08,836 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=460 eval_loss=0.7578 duration_sec=16.20
139
+ 2026-03-26 17:02:26,082 [INFO] new_opacus_codex.train_steps: epoch=3 step=465 loss=0.7848
140
+ 2026-03-26 17:02:43,254 [INFO] new_opacus_codex.train_steps: epoch=3 step=470 loss=0.7728
141
+ 2026-03-26 17:02:59,908 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=470 eval_loss=0.7577 duration_sec=16.62
142
+ 2026-03-26 17:03:16,979 [INFO] new_opacus_codex.train_steps: epoch=3 step=475 loss=0.7496
143
+ 2026-03-26 17:03:33,837 [INFO] new_opacus_codex.train_steps: epoch=3 step=480 loss=0.7514
144
+ 2026-03-26 17:03:49,969 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=480 eval_loss=0.7575 duration_sec=16.12
145
+ 2026-03-26 17:04:06,993 [INFO] new_opacus_codex.train_steps: epoch=3 step=485 loss=0.7676
146
+ 2026-03-26 17:04:24,097 [INFO] new_opacus_codex.train_steps: epoch=3 step=490 loss=0.7738
147
+ 2026-03-26 17:04:40,918 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=490 eval_loss=0.7574 duration_sec=16.80
148
+ 2026-03-26 17:04:58,307 [INFO] new_opacus_codex.train_steps: epoch=3 step=495 loss=0.7702
149
+ 2026-03-26 17:05:15,810 [INFO] new_opacus_codex.train_steps: epoch=3 step=500 loss=0.7781
150
+ 2026-03-26 17:05:32,238 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=500 eval_loss=0.7573 duration_sec=16.41
151
+ 2026-03-26 17:05:49,668 [INFO] new_opacus_codex.train_steps: epoch=3 step=505 loss=0.7648
152
+ 2026-03-26 17:06:07,107 [INFO] new_opacus_codex.train_steps: epoch=3 step=510 loss=0.7634
153
+ 2026-03-26 17:06:23,091 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=510 eval_loss=0.7572 duration_sec=15.97
154
+ 2026-03-26 17:06:40,297 [INFO] new_opacus_codex.train_steps: epoch=3 step=515 loss=0.7607
155
+ 2026-03-26 17:06:57,702 [INFO] new_opacus_codex.train_steps: epoch=3 step=520 loss=0.7721
156
+ 2026-03-26 17:07:14,222 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=520 eval_loss=0.7571 duration_sec=16.50
157
+ 2026-03-26 17:07:31,149 [INFO] new_opacus_codex.train_steps: epoch=3 step=525 loss=0.7681
158
+ 2026-03-26 17:07:48,232 [INFO] new_opacus_codex.train_steps: epoch=3 step=530 loss=0.7617
159
+ 2026-03-26 17:08:04,800 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=530 eval_loss=0.7571 duration_sec=16.55
160
+ 2026-03-26 17:08:21,499 [INFO] new_opacus_codex.train_steps: epoch=3 step=535 loss=0.7822
161
+ 2026-03-26 17:08:38,397 [INFO] new_opacus_codex.train_steps: epoch=3 step=540 loss=0.7775
162
+ 2026-03-26 17:08:55,120 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=540 eval_loss=0.7570 duration_sec=16.71
163
+ 2026-03-26 17:09:12,171 [INFO] new_opacus_codex.train_steps: epoch=3 step=545 loss=0.7652
164
+ 2026-03-26 17:09:29,442 [INFO] new_opacus_codex.train_steps: epoch=3 step=550 loss=0.7677
165
+ 2026-03-26 17:09:45,958 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=550 eval_loss=0.7570 duration_sec=16.50
166
+ 2026-03-26 17:10:04,549 [INFO] new_opacus_codex.train_steps: epoch=3 step=555 loss=0.8013
167
+ 2026-03-26 17:10:21,855 [INFO] new_opacus_codex.train_steps: epoch=3 step=560 loss=0.7772
168
+ 2026-03-26 17:10:37,678 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=560 eval_loss=0.7570 duration_sec=15.81
169
+ 2026-03-26 17:10:55,034 [INFO] new_opacus_codex.train_steps: epoch=3 step=565 loss=0.7598
170
+ 2026-03-26 17:11:12,062 [INFO] new_opacus_codex.train_steps: epoch=3 step=570 loss=0.7911
171
+ 2026-03-26 17:11:28,771 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=570 eval_loss=0.7569 duration_sec=16.70
172
+ 2026-03-26 17:11:45,671 [INFO] new_opacus_codex.train_steps: epoch=3 step=575 loss=0.7788
173
+ 2026-03-26 17:12:02,551 [INFO] new_opacus_codex.train_steps: epoch=3 step=580 loss=0.7568
174
+ 2026-03-26 17:12:19,071 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=580 eval_loss=0.7568 duration_sec=16.50
175
+ 2026-03-26 17:12:36,324 [INFO] new_opacus_codex.train_steps: epoch=3 step=585 loss=0.7475
176
+ 2026-03-26 17:12:53,426 [INFO] new_opacus_codex.train_steps: epoch=3 step=590 loss=0.7652
177
+ 2026-03-26 17:13:09,348 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=590 eval_loss=0.7568 duration_sec=15.90
178
+ 2026-03-26 17:13:26,183 [INFO] new_opacus_codex.train_steps: epoch=3 step=595 loss=0.7794
179
+ 2026-03-26 17:13:43,352 [INFO] new_opacus_codex.train_steps: epoch=3 step=600 loss=0.7577
180
+ 2026-03-26 17:14:00,653 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=600 eval_loss=0.7568 duration_sec=17.28
181
+ 2026-03-26 17:14:17,895 [INFO] new_opacus_codex.train_steps: epoch=3 step=605 loss=0.7630
182
+ 2026-03-26 17:14:34,934 [INFO] new_opacus_codex.train_steps: epoch=3 step=610 loss=0.7729
183
+ 2026-03-26 17:14:52,163 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=610 eval_loss=0.7567 duration_sec=17.22
184
+ 2026-03-26 17:15:09,096 [INFO] new_opacus_codex.train_steps: epoch=3 step=615 loss=0.7581
185
+ 2026-03-26 17:15:26,168 [INFO] new_opacus_codex.train_steps: epoch=3 step=620 loss=0.7678
186
+ 2026-03-26 17:15:42,813 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=620 eval_loss=0.7567 duration_sec=16.63
187
+ 2026-03-26 17:15:59,990 [INFO] new_opacus_codex.train_steps: epoch=3 step=625 loss=0.7859
188
+ 2026-03-26 17:16:17,253 [INFO] new_opacus_codex.train_steps: epoch=3 step=630 loss=0.7652
189
+ 2026-03-26 17:16:34,056 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=630 eval_loss=0.7567 duration_sec=16.79
190
+ 2026-03-26 17:16:51,431 [INFO] new_opacus_codex.train_steps: epoch=3 step=635 loss=0.7602
191
+ 2026-03-26 17:17:08,544 [INFO] new_opacus_codex.train_steps: epoch=3 step=640 loss=0.7757
192
+ 2026-03-26 17:17:24,994 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=640 eval_loss=0.7567 duration_sec=16.44
193
+ 2026-03-26 17:17:41,845 [INFO] new_opacus_codex.train_steps: epoch=3 step=645 loss=0.7567
194
+ 2026-03-26 17:17:58,914 [INFO] new_opacus_codex.train_steps: epoch=3 step=650 loss=0.7280
195
+ 2026-03-26 17:18:15,355 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=650 eval_loss=0.7567 duration_sec=16.43
granite-4.0-h-tiny/dp3_attn/adapter/README.md ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ibm-granite/granite-4.0-h-tiny-base
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:ibm-granite/granite-4.0-h-tiny-base
7
+ - lora
8
+ - transformers
9
+ ---
10
+
11
+ # Model Card for Model ID
12
+
13
+ <!-- Provide a quick summary of what the model is/does. -->
14
+
15
+
16
+
17
+ ## Model Details
18
+
19
+ ### Model Description
20
+
21
+ <!-- Provide a longer summary of what this model is. -->
22
+
23
+
24
+
25
+ - **Developed by:** [More Information Needed]
26
+ - **Funded by [optional]:** [More Information Needed]
27
+ - **Shared by [optional]:** [More Information Needed]
28
+ - **Model type:** [More Information Needed]
29
+ - **Language(s) (NLP):** [More Information Needed]
30
+ - **License:** [More Information Needed]
31
+ - **Finetuned from model [optional]:** [More Information Needed]
32
+
33
+ ### Model Sources [optional]
34
+
35
+ <!-- Provide the basic links for the model. -->
36
+
37
+ - **Repository:** [More Information Needed]
38
+ - **Paper [optional]:** [More Information Needed]
39
+ - **Demo [optional]:** [More Information Needed]
40
+
41
+ ## Uses
42
+
43
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
44
+
45
+ ### Direct Use
46
+
47
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
48
+
49
+ [More Information Needed]
50
+
51
+ ### Downstream Use [optional]
52
+
53
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
54
+
55
+ [More Information Needed]
56
+
57
+ ### Out-of-Scope Use
58
+
59
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
60
+
61
+ [More Information Needed]
62
+
63
+ ## Bias, Risks, and Limitations
64
+
65
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
66
+
67
+ [More Information Needed]
68
+
69
+ ### Recommendations
70
+
71
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
72
+
73
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
74
+
75
+ ## How to Get Started with the Model
76
+
77
+ Use the code below to get started with the model.
78
+
79
+ [More Information Needed]
80
+
81
+ ## Training Details
82
+
83
+ ### Training Data
84
+
85
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
86
+
87
+ [More Information Needed]
88
+
89
+ ### Training Procedure
90
+
91
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
92
+
93
+ #### Preprocessing [optional]
94
+
95
+ [More Information Needed]
96
+
97
+
98
+ #### Training Hyperparameters
99
+
100
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
101
+
102
+ #### Speeds, Sizes, Times [optional]
103
+
104
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
105
+
106
+ [More Information Needed]
107
+
108
+ ## Evaluation
109
+
110
+ <!-- This section describes the evaluation protocols and provides the results. -->
111
+
112
+ ### Testing Data, Factors & Metrics
113
+
114
+ #### Testing Data
115
+
116
+ <!-- This should link to a Dataset Card if possible. -->
117
+
118
+ [More Information Needed]
119
+
120
+ #### Factors
121
+
122
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
123
+
124
+ [More Information Needed]
125
+
126
+ #### Metrics
127
+
128
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
129
+
130
+ [More Information Needed]
131
+
132
+ ### Results
133
+
134
+ [More Information Needed]
135
+
136
+ #### Summary
137
+
138
+
139
+
140
+ ## Model Examination [optional]
141
+
142
+ <!-- Relevant interpretability work for the model goes here -->
143
+
144
+ [More Information Needed]
145
+
146
+ ## Environmental Impact
147
+
148
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
149
+
150
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
151
+
152
+ - **Hardware Type:** [More Information Needed]
153
+ - **Hours used:** [More Information Needed]
154
+ - **Cloud Provider:** [More Information Needed]
155
+ - **Compute Region:** [More Information Needed]
156
+ - **Carbon Emitted:** [More Information Needed]
157
+
158
+ ## Technical Specifications [optional]
159
+
160
+ ### Model Architecture and Objective
161
+
162
+ [More Information Needed]
163
+
164
+ ### Compute Infrastructure
165
+
166
+ [More Information Needed]
167
+
168
+ #### Hardware
169
+
170
+ [More Information Needed]
171
+
172
+ #### Software
173
+
174
+ [More Information Needed]
175
+
176
+ ## Citation [optional]
177
+
178
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
179
+
180
+ **BibTeX:**
181
+
182
+ [More Information Needed]
183
+
184
+ **APA:**
185
+
186
+ [More Information Needed]
187
+
188
+ ## Glossary [optional]
189
+
190
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
191
+
192
+ [More Information Needed]
193
+
194
+ ## More Information [optional]
195
+
196
+ [More Information Needed]
197
+
198
+ ## Model Card Authors [optional]
199
+
200
+ [More Information Needed]
201
+
202
+ ## Model Card Contact
203
+
204
+ [More Information Needed]
205
+ ### Framework versions
206
+
207
+ - PEFT 0.18.1
granite-4.0-h-tiny/dp3_attn/adapter/adapter_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "ibm-granite/granite-4.0-h-tiny-base",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": true,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": [
25
+ "lm_head",
26
+ "embed_tokens"
27
+ ],
28
+ "peft_type": "LORA",
29
+ "peft_version": "0.18.1",
30
+ "qalora_group_size": 16,
31
+ "r": 16,
32
+ "rank_pattern": {},
33
+ "revision": null,
34
+ "target_modules": [
35
+ "v_proj",
36
+ "o_proj",
37
+ "k_proj",
38
+ "q_proj"
39
+ ],
40
+ "target_parameters": null,
41
+ "task_type": "CAUSAL_LM",
42
+ "trainable_token_indices": null,
43
+ "use_dora": false,
44
+ "use_qalora": false,
45
+ "use_rslora": false
46
+ }
granite-4.0-h-tiny/dp3_attn/audit_results.json ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "delta": 1e-05,
3
+ "num_canaries": 500,
4
+ "num_members": 250,
5
+ "paper_guess_fraction": 0.2,
6
+ "paper_guess_steps": 20,
7
+ "loss": {
8
+ "auc": 0.547944,
9
+ "empirical_epsilon": {
10
+ "0.05": 0.14375991560518742,
11
+ "0.01": 0.0
12
+ },
13
+ "empirical_epsilon_details": {
14
+ "0.05": {
15
+ "epsilon": 0.14375991560518742,
16
+ "num_guesses": 60,
17
+ "correct_guesses": 39,
18
+ "candidate_num_guesses": [
19
+ 5,
20
+ 10,
21
+ 15,
22
+ 20,
23
+ 25,
24
+ 30,
25
+ 35,
26
+ 40,
27
+ 45,
28
+ 50,
29
+ 55,
30
+ 60,
31
+ 65,
32
+ 70,
33
+ 75,
34
+ 80,
35
+ 85,
36
+ 90,
37
+ 95,
38
+ 100
39
+ ],
40
+ "direction": "lower"
41
+ },
42
+ "0.01": {
43
+ "epsilon": 0.0,
44
+ "num_guesses": 0,
45
+ "correct_guesses": 0,
46
+ "candidate_num_guesses": [
47
+ 5,
48
+ 10,
49
+ 15,
50
+ 20,
51
+ 25,
52
+ 30,
53
+ 35,
54
+ 40,
55
+ 45,
56
+ 50,
57
+ 55,
58
+ 60,
59
+ 65,
60
+ 70,
61
+ 75,
62
+ 80,
63
+ 85,
64
+ 90,
65
+ 95,
66
+ 100
67
+ ],
68
+ "direction": "lower"
69
+ }
70
+ }
71
+ },
72
+ "embedding": {
73
+ "auc": 0.527048,
74
+ "empirical_epsilon": {
75
+ "0.05": 0.0,
76
+ "0.01": 0.0
77
+ },
78
+ "empirical_epsilon_details": {
79
+ "0.05": {
80
+ "epsilon": 0.0,
81
+ "num_guesses": 0,
82
+ "correct_guesses": 0,
83
+ "candidate_num_guesses": [
84
+ 5,
85
+ 10,
86
+ 15,
87
+ 20,
88
+ 25,
89
+ 30,
90
+ 35,
91
+ 40,
92
+ 45,
93
+ 50,
94
+ 55,
95
+ 60,
96
+ 65,
97
+ 70,
98
+ 75,
99
+ 80,
100
+ 85,
101
+ 90,
102
+ 95,
103
+ 100
104
+ ],
105
+ "direction": "lower"
106
+ },
107
+ "0.01": {
108
+ "epsilon": 0.0,
109
+ "num_guesses": 0,
110
+ "correct_guesses": 0,
111
+ "candidate_num_guesses": [
112
+ 5,
113
+ 10,
114
+ 15,
115
+ 20,
116
+ 25,
117
+ 30,
118
+ 35,
119
+ 40,
120
+ 45,
121
+ 50,
122
+ 55,
123
+ 60,
124
+ 65,
125
+ 70,
126
+ 75,
127
+ 80,
128
+ 85,
129
+ 90,
130
+ 95,
131
+ 100
132
+ ],
133
+ "direction": "lower"
134
+ }
135
+ }
136
+ }
137
+ }
granite-4.0-h-tiny/dp3_attn/canary_meta.json ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/dp3_attn/metrics.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/dp3_attn/resolved_config.yaml ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ name: ibm-granite/granite-4.0-h-tiny-base
3
+ tokenizer_name: ibm-granite/granite-4.0-h-tiny-base
4
+ max_length: 1024
5
+ dtype: bfloat16
6
+ trust_remote_code: true
7
+ use_fast_tokenizer: true
8
+ cache_dir: null
9
+ local_files_only: false
10
+ low_cpu_mem_usage: true
11
+ tie_word_embeddings: true
12
+ gradient_checkpointing: false
13
+ use_chat_template: false
14
+ dataset:
15
+ name: melihcatal/codedp-cpt
16
+ split: train
17
+ mode: cpt
18
+ text_column: text
19
+ validation_ratio: 0.05
20
+ max_samples: -1
21
+ lora:
22
+ enabled: true
23
+ r: 16
24
+ alpha: 32
25
+ dropout: 0.05
26
+ target_modules:
27
+ - q_proj
28
+ - k_proj
29
+ - v_proj
30
+ - o_proj
31
+ modules_to_save:
32
+ - lm_head
33
+ bias: none
34
+ training:
35
+ seed: 42
36
+ epochs: 3
37
+ warmup_steps: null
38
+ warmup_ratio: 0.05
39
+ mixed_precision: false
40
+ mixed_precision_dtype: bfloat16
41
+ batch_size: 8
42
+ eval_batch_size: 8
43
+ eval_every_steps: 10
44
+ eval_every_epochs: 1
45
+ learning_rate: 0.00017
46
+ optimizer: adamw
47
+ lr_scheduler: cosine
48
+ adam_beta1: 0.9
49
+ adam_beta2: 0.999
50
+ adam_epsilon: 1.0e-08
51
+ sgd_momentum: 0.9
52
+ weight_decay: 0.01
53
+ max_grad_norm: 1.0
54
+ log_every: 5
55
+ gradient_accumulation_steps: 4
56
+ num_workers: 4
57
+ output_dir: /scratch/mcatal/runs/cpt/granite-4.0-h-tiny/dp3_attn
58
+ distributed:
59
+ strategy: dpddp
60
+ backend: nccl
61
+ devices: null
62
+ dp:
63
+ module_validator: auto
64
+ target_delta: 1.0e-05
65
+ noise_multiplier: null
66
+ max_grad_norm: 1.0
67
+ grad_sample_mode: hooks
68
+ clipping: flat
69
+ secure_mode: false
70
+ enabled: true
71
+ target_epsilon: 3.0
72
+ audit:
73
+ enabled: true
74
+ run_every_epoch: true
75
+ epoch_device: cuda
76
+ q_canary: auto
77
+ num_canaries: 500
78
+ prefix_length: 49
79
+ num_digits: 12
80
+ batch_size: 32
81
+ delta: 1.0e-05
82
+ p_values:
83
+ - 0.05
84
+ - 0.01
85
+ paper_guess_fraction: 0.2
86
+ paper_guess_steps: 20
87
+ enable_holdout_empirical_epsilon: false
88
+ holdout_seed: 42
89
+ tie_seed: 42
90
+ tracking:
91
+ enabled: true
92
+ tensorboard: true
93
+ wandb: false
94
+ wandb_project: codedp-finetune-h200-audit
95
+ wandb_run_name: granite-4.0-h-tiny-cpt-dp3_attn
96
+ wandb_mode: online
97
+ codecarbon: true
98
+ codecarbon_output_file: codecarbon.csv
99
+ codecarbon_measure_power_secs: 15
100
+ codecarbon_country_iso_code: null
101
+ codecarbon_project_name: codedp-granite-4.0-h-tiny-cpt-dp3_attn
102
+ moe:
103
+ output_router_logits: false
104
+ router_aux_loss_coef: 0.0
105
+ freeze_router: true
106
+ profile:
107
+ enabled: false
108
+ num_batches: 8
109
+ top_experts: 8
110
+ output_file: moe_expert_profile.json
granite-4.0-h-tiny/dp3_attn/summary.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audit/delta": 1e-05,
3
+ "audit/embedding/auc": 0.527048,
4
+ "audit/embedding/empirical_epsilon/0.01": 0.0,
5
+ "audit/embedding/empirical_epsilon/0.05": 0.0,
6
+ "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 0.0,
7
+ "audit/embedding/empirical_epsilon_details/0.01/epsilon": 0.0,
8
+ "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 0.0,
9
+ "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 0.0,
10
+ "audit/embedding/empirical_epsilon_details/0.05/epsilon": 0.0,
11
+ "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 0.0,
12
+ "audit/loss/auc": 0.547944,
13
+ "audit/loss/empirical_epsilon/0.01": 0.0,
14
+ "audit/loss/empirical_epsilon/0.05": 0.14375991560518742,
15
+ "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 0.0,
16
+ "audit/loss/empirical_epsilon_details/0.01/epsilon": 0.0,
17
+ "audit/loss/empirical_epsilon_details/0.01/num_guesses": 0.0,
18
+ "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 39.0,
19
+ "audit/loss/empirical_epsilon_details/0.05/epsilon": 0.14375991560518742,
20
+ "audit/loss/empirical_epsilon_details/0.05/num_guesses": 60.0,
21
+ "audit/num_canaries": 500.0,
22
+ "audit/num_members": 250.0,
23
+ "audit/paper_guess_fraction": 0.2,
24
+ "audit/paper_guess_steps": 20.0,
25
+ "energy/codecarbon/cpu_count": 8.0,
26
+ "energy/codecarbon/cpu_energy": 0.11020718112665469,
27
+ "energy/codecarbon/cpu_power": 80.65461212393382,
28
+ "energy/codecarbon/cpu_utilization_percent": 9.477919779787651,
29
+ "energy/codecarbon/duration": 5116.045043497812,
30
+ "energy/codecarbon/emissions": 0.11511925962159554,
31
+ "energy/codecarbon/emissions_rate": 2.2501611819838306e-05,
32
+ "energy/codecarbon/energy_consumed": 3.3039422443990336,
33
+ "energy/codecarbon/gpu_count": 8.0,
34
+ "energy/codecarbon/gpu_energy": 3.1418119734475454,
35
+ "energy/codecarbon/gpu_power": 2221.8516828513934,
36
+ "energy/codecarbon/gpu_utilization_percent": 65.41046500196619,
37
+ "energy/codecarbon/latitude": 47.4843,
38
+ "energy/codecarbon/longitude": 8.212,
39
+ "energy/codecarbon/pue": 1.0,
40
+ "energy/codecarbon/ram_energy": 0.05192308982483644,
41
+ "energy/codecarbon/ram_power": 38.0,
42
+ "energy/codecarbon/ram_total_size": 256.0,
43
+ "energy/codecarbon/ram_used_gb": 528.4883023163394,
44
+ "energy/codecarbon/ram_utilization_percent": 26.660813999213527,
45
+ "energy/codecarbon/water_consumed": 0.0,
46
+ "energy/codecarbon/wue": 0.0,
47
+ "eval/duration_sec": 17.514180964091793,
48
+ "eval/loss": 1.053442443518535,
49
+ "perf/audit_duration_sec": 7.912764948094264,
50
+ "perf/epoch_duration_sec": 1698.1409097390715,
51
+ "perf/epoch_samples": 61331.0,
52
+ "perf/epoch_samples_per_sec": 36.116555256550434,
53
+ "perf/epoch_tokens": 47090629.0,
54
+ "perf/epoch_tokens_per_sec": 27730.69580382215,
55
+ "perf/logical_batch_size": 35.0,
56
+ "perf/logical_token_count": 28696.0,
57
+ "perf/physical_batches": 5.0,
58
+ "perf/samples_per_sec": 5.637019105396721,
59
+ "perf/step_duration_sec": 6.208955361973494,
60
+ "perf/tokens_per_sec": 4621.711435670409,
61
+ "privacy/epsilon": 2.99107185013027,
62
+ "system/cuda_epoch_peak_memory_gb": 54.1029486656189,
63
+ "system/cuda_max_memory_allocated_gb": 54.1029486656189,
64
+ "system/cuda_memory_allocated_gb": 15.311615943908691,
65
+ "train/epoch_canary_loss": 14.617763124949052,
66
+ "train/epoch_loss": 2.397428680087968,
67
+ "train/epoch_real_loss": 1.1449934101948456,
68
+ "train/lr": 4.308993292072017e-09,
69
+ "train/step_canary_loss": 14.5,
70
+ "train/step_loss": 2.2486216136387416,
71
+ "train/step_real_loss": 1.1000548899173737
72
+ }
granite-4.0-h-tiny/dp3_attn/tokenizer/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/dp3_attn/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,516 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|end_of_text|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|end_of_text|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": {
9
+ "extra_special_token_0": "865331112869",
10
+ "extra_special_token_1": "569765693871",
11
+ "extra_special_token_2": "485177821815",
12
+ "extra_special_token_3": "135441121756",
13
+ "extra_special_token_4": "367459894796",
14
+ "extra_special_token_5": "877482678543",
15
+ "extra_special_token_6": "457919547633",
16
+ "extra_special_token_7": "765474393376",
17
+ "extra_special_token_8": "114848338811",
18
+ "extra_special_token_9": "746285987371",
19
+ "extra_special_token_10": "649291669397",
20
+ "extra_special_token_11": "927914615679",
21
+ "extra_special_token_12": "445925149649",
22
+ "extra_special_token_13": "691587454538",
23
+ "extra_special_token_14": "143777992227",
24
+ "extra_special_token_15": "997981281989",
25
+ "extra_special_token_16": "425949483533",
26
+ "extra_special_token_17": "982993456429",
27
+ "extra_special_token_18": "718726519731",
28
+ "extra_special_token_19": "172599315861",
29
+ "extra_special_token_20": "643489267333",
30
+ "extra_special_token_21": "282322838685",
31
+ "extra_special_token_22": "781653545886",
32
+ "extra_special_token_23": "796415361892",
33
+ "extra_special_token_24": "841991688488",
34
+ "extra_special_token_25": "211411365397",
35
+ "extra_special_token_26": "698218415444",
36
+ "extra_special_token_27": "355977139358",
37
+ "extra_special_token_28": "682564697312",
38
+ "extra_special_token_29": "383837596997",
39
+ "extra_special_token_30": "689362171782",
40
+ "extra_special_token_31": "749966767285",
41
+ "extra_special_token_32": "753159165157",
42
+ "extra_special_token_33": "795693824762",
43
+ "extra_special_token_34": "669689115557",
44
+ "extra_special_token_35": "327491773134",
45
+ "extra_special_token_36": "983569279932",
46
+ "extra_special_token_37": "612128769512",
47
+ "extra_special_token_38": "374327157578",
48
+ "extra_special_token_39": "311632789559",
49
+ "extra_special_token_40": "523918658846",
50
+ "extra_special_token_41": "765981581453",
51
+ "extra_special_token_42": "794825141891",
52
+ "extra_special_token_43": "873898736873",
53
+ "extra_special_token_44": "447445629421",
54
+ "extra_special_token_45": "473822473819",
55
+ "extra_special_token_46": "181439694557",
56
+ "extra_special_token_47": "592538279337",
57
+ "extra_special_token_48": "668134915514",
58
+ "extra_special_token_49": "643692393748",
59
+ "extra_special_token_50": "696651276628",
60
+ "extra_special_token_51": "853859348234",
61
+ "extra_special_token_52": "778466723723",
62
+ "extra_special_token_53": "929826356991",
63
+ "extra_special_token_54": "272362973463",
64
+ "extra_special_token_55": "694235616268",
65
+ "extra_special_token_56": "281673864127",
66
+ "extra_special_token_57": "479676316326",
67
+ "extra_special_token_58": "646979124677",
68
+ "extra_special_token_59": "922327493433",
69
+ "extra_special_token_60": "883685933161",
70
+ "extra_special_token_61": "264259917554",
71
+ "extra_special_token_62": "836746273134",
72
+ "extra_special_token_63": "658481324922",
73
+ "extra_special_token_64": "481884157827",
74
+ "extra_special_token_65": "587787496812",
75
+ "extra_special_token_66": "579184949249",
76
+ "extra_special_token_67": "912193598348",
77
+ "extra_special_token_68": "529679678956",
78
+ "extra_special_token_69": "795838284624",
79
+ "extra_special_token_70": "159337222655",
80
+ "extra_special_token_71": "173781362446",
81
+ "extra_special_token_72": "773687856563",
82
+ "extra_special_token_73": "535787224917",
83
+ "extra_special_token_74": "351885857332",
84
+ "extra_special_token_75": "578827344666",
85
+ "extra_special_token_76": "198462689911",
86
+ "extra_special_token_77": "722618266242",
87
+ "extra_special_token_78": "952872416512",
88
+ "extra_special_token_79": "517778845323",
89
+ "extra_special_token_80": "749665846687",
90
+ "extra_special_token_81": "661436365453",
91
+ "extra_special_token_82": "259666844669",
92
+ "extra_special_token_83": "242851284913",
93
+ "extra_special_token_84": "514532995959",
94
+ "extra_special_token_85": "161588262349",
95
+ "extra_special_token_86": "742765629356",
96
+ "extra_special_token_87": "225164373623",
97
+ "extra_special_token_88": "676539973863",
98
+ "extra_special_token_89": "826214551218",
99
+ "extra_special_token_90": "182345464792",
100
+ "extra_special_token_91": "232776999554",
101
+ "extra_special_token_92": "337326533813",
102
+ "extra_special_token_93": "676676697292",
103
+ "extra_special_token_94": "929185622831",
104
+ "extra_special_token_95": "545512344383",
105
+ "extra_special_token_96": "499444466686",
106
+ "extra_special_token_97": "314697386682",
107
+ "extra_special_token_98": "517379856925",
108
+ "extra_special_token_99": "379557332953",
109
+ "extra_special_token_100": "614797267726",
110
+ "extra_special_token_101": "429781429464",
111
+ "extra_special_token_102": "922466849763",
112
+ "extra_special_token_103": "721737645236",
113
+ "extra_special_token_104": "479227349997",
114
+ "extra_special_token_105": "136931728327",
115
+ "extra_special_token_106": "259533577263",
116
+ "extra_special_token_107": "488538864842",
117
+ "extra_special_token_108": "937495658852",
118
+ "extra_special_token_109": "489991411364",
119
+ "extra_special_token_110": "499148455254",
120
+ "extra_special_token_111": "441373944925",
121
+ "extra_special_token_112": "899151413682",
122
+ "extra_special_token_113": "467893531755",
123
+ "extra_special_token_114": "527117488925",
124
+ "extra_special_token_115": "928335588653",
125
+ "extra_special_token_116": "374439448821",
126
+ "extra_special_token_117": "879425227932",
127
+ "extra_special_token_118": "867678158885",
128
+ "extra_special_token_119": "399749397872",
129
+ "extra_special_token_120": "129693547287",
130
+ "extra_special_token_121": "689285841825",
131
+ "extra_special_token_122": "771619544974",
132
+ "extra_special_token_123": "724883568652",
133
+ "extra_special_token_124": "516968424863",
134
+ "extra_special_token_125": "733737988257",
135
+ "extra_special_token_126": "852347289392",
136
+ "extra_special_token_127": "296953381169",
137
+ "extra_special_token_128": "377273562477",
138
+ "extra_special_token_129": "262296912232",
139
+ "extra_special_token_130": "547149832394",
140
+ "extra_special_token_131": "298464134954",
141
+ "extra_special_token_132": "216667245274",
142
+ "extra_special_token_133": "843998562287",
143
+ "extra_special_token_134": "572154333646",
144
+ "extra_special_token_135": "124589118494",
145
+ "extra_special_token_136": "841824384614",
146
+ "extra_special_token_137": "232896526252",
147
+ "extra_special_token_138": "295448593321",
148
+ "extra_special_token_139": "123741461297",
149
+ "extra_special_token_140": "653573457168",
150
+ "extra_special_token_141": "196735786156",
151
+ "extra_special_token_142": "377338713663",
152
+ "extra_special_token_143": "964342468552",
153
+ "extra_special_token_144": "586855179568",
154
+ "extra_special_token_145": "484773717614",
155
+ "extra_special_token_146": "894885246797",
156
+ "extra_special_token_147": "677896358599",
157
+ "extra_special_token_148": "848845611563",
158
+ "extra_special_token_149": "851852651677",
159
+ "extra_special_token_150": "398549545767",
160
+ "extra_special_token_151": "454244839926",
161
+ "extra_special_token_152": "799364566435",
162
+ "extra_special_token_153": "967114116556",
163
+ "extra_special_token_154": "817378986438",
164
+ "extra_special_token_155": "233795848681",
165
+ "extra_special_token_156": "824387273757",
166
+ "extra_special_token_157": "916198946615",
167
+ "extra_special_token_158": "563117729724",
168
+ "extra_special_token_159": "951794811935",
169
+ "extra_special_token_160": "374598961236",
170
+ "extra_special_token_161": "922867396683",
171
+ "extra_special_token_162": "765737843639",
172
+ "extra_special_token_163": "175469284871",
173
+ "extra_special_token_164": "231853711778",
174
+ "extra_special_token_165": "662426712668",
175
+ "extra_special_token_166": "711412347158",
176
+ "extra_special_token_167": "753466987363",
177
+ "extra_special_token_168": "513361312532",
178
+ "extra_special_token_169": "712992815957",
179
+ "extra_special_token_170": "971621888444",
180
+ "extra_special_token_171": "829235161526",
181
+ "extra_special_token_172": "585544633356",
182
+ "extra_special_token_173": "582471228164",
183
+ "extra_special_token_174": "678666359123",
184
+ "extra_special_token_175": "557533689478",
185
+ "extra_special_token_176": "632962475133",
186
+ "extra_special_token_177": "484489193824",
187
+ "extra_special_token_178": "489562189822",
188
+ "extra_special_token_179": "589547936288",
189
+ "extra_special_token_180": "363214487524",
190
+ "extra_special_token_181": "244885399387",
191
+ "extra_special_token_182": "431751228368",
192
+ "extra_special_token_183": "433581868192",
193
+ "extra_special_token_184": "486391569221",
194
+ "extra_special_token_185": "185438575221",
195
+ "extra_special_token_186": "126574388585",
196
+ "extra_special_token_187": "741757479784",
197
+ "extra_special_token_188": "529854679937",
198
+ "extra_special_token_189": "996116119839",
199
+ "extra_special_token_190": "616248973917",
200
+ "extra_special_token_191": "763531783491",
201
+ "extra_special_token_192": "955456118295",
202
+ "extra_special_token_193": "364196983365",
203
+ "extra_special_token_194": "195792996468",
204
+ "extra_special_token_195": "151859598873",
205
+ "extra_special_token_196": "399223169721",
206
+ "extra_special_token_197": "938488813964",
207
+ "extra_special_token_198": "961981959227",
208
+ "extra_special_token_199": "183368827562",
209
+ "extra_special_token_200": "533417736566",
210
+ "extra_special_token_201": "786391632558",
211
+ "extra_special_token_202": "665661658354",
212
+ "extra_special_token_203": "693281533643",
213
+ "extra_special_token_204": "475794684356",
214
+ "extra_special_token_205": "652154162978",
215
+ "extra_special_token_206": "753233719644",
216
+ "extra_special_token_207": "668514843129",
217
+ "extra_special_token_208": "819162623892",
218
+ "extra_special_token_209": "941169431859",
219
+ "extra_special_token_210": "877385381798",
220
+ "extra_special_token_211": "752644929761",
221
+ "extra_special_token_212": "881136466196",
222
+ "extra_special_token_213": "275597777299",
223
+ "extra_special_token_214": "731681792655",
224
+ "extra_special_token_215": "961133895172",
225
+ "extra_special_token_216": "864718285734",
226
+ "extra_special_token_217": "963852916563",
227
+ "extra_special_token_218": "319584985416",
228
+ "extra_special_token_219": "563365646341",
229
+ "extra_special_token_220": "811371928234",
230
+ "extra_special_token_221": "837131396371",
231
+ "extra_special_token_222": "267514771964",
232
+ "extra_special_token_223": "944513428457",
233
+ "extra_special_token_224": "117298239631",
234
+ "extra_special_token_225": "158142752582",
235
+ "extra_special_token_226": "252867443568",
236
+ "extra_special_token_227": "839269684865",
237
+ "extra_special_token_228": "612788593128",
238
+ "extra_special_token_229": "145669731981",
239
+ "extra_special_token_230": "121557291859",
240
+ "extra_special_token_231": "245416776926",
241
+ "extra_special_token_232": "799417897197",
242
+ "extra_special_token_233": "997958836435",
243
+ "extra_special_token_234": "892336777248",
244
+ "extra_special_token_235": "158929292238",
245
+ "extra_special_token_236": "581976444672",
246
+ "extra_special_token_237": "897784492783",
247
+ "extra_special_token_238": "492373714791",
248
+ "extra_special_token_239": "512659818733",
249
+ "extra_special_token_240": "881112998642",
250
+ "extra_special_token_241": "619454958782",
251
+ "extra_special_token_242": "431149748713",
252
+ "extra_special_token_243": "624221476921",
253
+ "extra_special_token_244": "125866399464",
254
+ "extra_special_token_245": "339882449689",
255
+ "extra_special_token_246": "186198784585",
256
+ "extra_special_token_247": "943193294691",
257
+ "extra_special_token_248": "955668961269",
258
+ "extra_special_token_249": "232787996724",
259
+ "extra_special_token_250": "215671314196",
260
+ "extra_special_token_251": "286173241916",
261
+ "extra_special_token_252": "745977673725",
262
+ "extra_special_token_253": "556976448182",
263
+ "extra_special_token_254": "599961512792",
264
+ "extra_special_token_255": "766294538337",
265
+ "extra_special_token_256": "934912591213",
266
+ "extra_special_token_257": "295118729589",
267
+ "extra_special_token_258": "529455466433",
268
+ "extra_special_token_259": "196119929397",
269
+ "extra_special_token_260": "379571934299",
270
+ "extra_special_token_261": "251789649997",
271
+ "extra_special_token_262": "564544131355",
272
+ "extra_special_token_263": "244371196654",
273
+ "extra_special_token_264": "384598329253",
274
+ "extra_special_token_265": "887753195844",
275
+ "extra_special_token_266": "364947325679",
276
+ "extra_special_token_267": "655517954651",
277
+ "extra_special_token_268": "673948786567",
278
+ "extra_special_token_269": "857231548835",
279
+ "extra_special_token_270": "816115936673",
280
+ "extra_special_token_271": "644234165531",
281
+ "extra_special_token_272": "182782912224",
282
+ "extra_special_token_273": "234316622259",
283
+ "extra_special_token_274": "421369185549",
284
+ "extra_special_token_275": "434632855397",
285
+ "extra_special_token_276": "921889371893",
286
+ "extra_special_token_277": "415956914763",
287
+ "extra_special_token_278": "598916996413",
288
+ "extra_special_token_279": "773671349113",
289
+ "extra_special_token_280": "952465217972",
290
+ "extra_special_token_281": "117657531962",
291
+ "extra_special_token_282": "729825168745",
292
+ "extra_special_token_283": "691315125346",
293
+ "extra_special_token_284": "768461952319",
294
+ "extra_special_token_285": "664847713559",
295
+ "extra_special_token_286": "953267689786",
296
+ "extra_special_token_287": "886464195129",
297
+ "extra_special_token_288": "824488329416",
298
+ "extra_special_token_289": "837873762491",
299
+ "extra_special_token_290": "532833541879",
300
+ "extra_special_token_291": "669183782449",
301
+ "extra_special_token_292": "941976537588",
302
+ "extra_special_token_293": "739394546916",
303
+ "extra_special_token_294": "267954879268",
304
+ "extra_special_token_295": "637551427887",
305
+ "extra_special_token_296": "217756494954",
306
+ "extra_special_token_297": "524444658383",
307
+ "extra_special_token_298": "117783274348",
308
+ "extra_special_token_299": "138218735276",
309
+ "extra_special_token_300": "814611949491",
310
+ "extra_special_token_301": "711641973413",
311
+ "extra_special_token_302": "499156317423",
312
+ "extra_special_token_303": "515856611931",
313
+ "extra_special_token_304": "454164859837",
314
+ "extra_special_token_305": "345271433112",
315
+ "extra_special_token_306": "462294118988",
316
+ "extra_special_token_307": "511785788222",
317
+ "extra_special_token_308": "497294727353",
318
+ "extra_special_token_309": "866519986723",
319
+ "extra_special_token_310": "334513529294",
320
+ "extra_special_token_311": "549946382131",
321
+ "extra_special_token_312": "284445431422",
322
+ "extra_special_token_313": "396521188476",
323
+ "extra_special_token_314": "421435255895",
324
+ "extra_special_token_315": "133373659361",
325
+ "extra_special_token_316": "322683334381",
326
+ "extra_special_token_317": "228358422847",
327
+ "extra_special_token_318": "291762694874",
328
+ "extra_special_token_319": "143182978129",
329
+ "extra_special_token_320": "511923256573",
330
+ "extra_special_token_321": "327158398268",
331
+ "extra_special_token_322": "879764613759",
332
+ "extra_special_token_323": "564395222747",
333
+ "extra_special_token_324": "451161679736",
334
+ "extra_special_token_325": "538631466654",
335
+ "extra_special_token_326": "221762325616",
336
+ "extra_special_token_327": "218391991184",
337
+ "extra_special_token_328": "322589379462",
338
+ "extra_special_token_329": "876537814263",
339
+ "extra_special_token_330": "152676556624",
340
+ "extra_special_token_331": "332522971941",
341
+ "extra_special_token_332": "884354318946",
342
+ "extra_special_token_333": "513349618943",
343
+ "extra_special_token_334": "116639746413",
344
+ "extra_special_token_335": "635185846287",
345
+ "extra_special_token_336": "993832498489",
346
+ "extra_special_token_337": "813981174797",
347
+ "extra_special_token_338": "438745114173",
348
+ "extra_special_token_339": "983493951323",
349
+ "extra_special_token_340": "724492262421",
350
+ "extra_special_token_341": "622553389126",
351
+ "extra_special_token_342": "889965243135",
352
+ "extra_special_token_343": "364492359246",
353
+ "extra_special_token_344": "154962668224",
354
+ "extra_special_token_345": "179564995814",
355
+ "extra_special_token_346": "418412875665",
356
+ "extra_special_token_347": "718951851413",
357
+ "extra_special_token_348": "699446724178",
358
+ "extra_special_token_349": "624266421831",
359
+ "extra_special_token_350": "815458725125",
360
+ "extra_special_token_351": "455423278865",
361
+ "extra_special_token_352": "393741199486",
362
+ "extra_special_token_353": "328552864359",
363
+ "extra_special_token_354": "211662639865",
364
+ "extra_special_token_355": "218784516525",
365
+ "extra_special_token_356": "762486672996",
366
+ "extra_special_token_357": "142799718159",
367
+ "extra_special_token_358": "858146415154",
368
+ "extra_special_token_359": "767858144912",
369
+ "extra_special_token_360": "571317457151",
370
+ "extra_special_token_361": "635127952696",
371
+ "extra_special_token_362": "116427191984",
372
+ "extra_special_token_363": "268921994538",
373
+ "extra_special_token_364": "523937669294",
374
+ "extra_special_token_365": "165429152138",
375
+ "extra_special_token_366": "739246183345",
376
+ "extra_special_token_367": "591464355756",
377
+ "extra_special_token_368": "212985874612",
378
+ "extra_special_token_369": "191887635211",
379
+ "extra_special_token_370": "967214577653",
380
+ "extra_special_token_371": "119342152414",
381
+ "extra_special_token_372": "946444632795",
382
+ "extra_special_token_373": "618423867817",
383
+ "extra_special_token_374": "228565148417",
384
+ "extra_special_token_375": "729116422489",
385
+ "extra_special_token_376": "527874729936",
386
+ "extra_special_token_377": "739784153482",
387
+ "extra_special_token_378": "387763951128",
388
+ "extra_special_token_379": "331369926711",
389
+ "extra_special_token_380": "562716493614",
390
+ "extra_special_token_381": "739667844957",
391
+ "extra_special_token_382": "562389434565",
392
+ "extra_special_token_383": "256497188281",
393
+ "extra_special_token_384": "859927364588",
394
+ "extra_special_token_385": "417668946583",
395
+ "extra_special_token_386": "357621613582",
396
+ "extra_special_token_387": "438435178228",
397
+ "extra_special_token_388": "485692541169",
398
+ "extra_special_token_389": "825815739116",
399
+ "extra_special_token_390": "342221452223",
400
+ "extra_special_token_391": "697747991249",
401
+ "extra_special_token_392": "716763689965",
402
+ "extra_special_token_393": "141499982867",
403
+ "extra_special_token_394": "818479319499",
404
+ "extra_special_token_395": "336813343298",
405
+ "extra_special_token_396": "594688742928",
406
+ "extra_special_token_397": "472129283475",
407
+ "extra_special_token_398": "514354144759",
408
+ "extra_special_token_399": "349249721685",
409
+ "extra_special_token_400": "546276298359",
410
+ "extra_special_token_401": "353755529131",
411
+ "extra_special_token_402": "315534574435",
412
+ "extra_special_token_403": "523723475786",
413
+ "extra_special_token_404": "215826764872",
414
+ "extra_special_token_405": "367968398551",
415
+ "extra_special_token_406": "569853653352",
416
+ "extra_special_token_407": "389715484387",
417
+ "extra_special_token_408": "293847485454",
418
+ "extra_special_token_409": "714738141818",
419
+ "extra_special_token_410": "178478368922",
420
+ "extra_special_token_411": "581493616981",
421
+ "extra_special_token_412": "589439538674",
422
+ "extra_special_token_413": "846657726193",
423
+ "extra_special_token_414": "722339992679",
424
+ "extra_special_token_415": "138154781148",
425
+ "extra_special_token_416": "757785319772",
426
+ "extra_special_token_417": "492516914298",
427
+ "extra_special_token_418": "919181521716",
428
+ "extra_special_token_419": "985781138935",
429
+ "extra_special_token_420": "476969195485",
430
+ "extra_special_token_421": "313145133463",
431
+ "extra_special_token_422": "758963111966",
432
+ "extra_special_token_423": "147541537162",
433
+ "extra_special_token_424": "557163366873",
434
+ "extra_special_token_425": "144373897488",
435
+ "extra_special_token_426": "522515164754",
436
+ "extra_special_token_427": "724964923582",
437
+ "extra_special_token_428": "284776712475",
438
+ "extra_special_token_429": "375429755114",
439
+ "extra_special_token_430": "181233596124",
440
+ "extra_special_token_431": "948585673431",
441
+ "extra_special_token_432": "243165586174",
442
+ "extra_special_token_433": "396847976144",
443
+ "extra_special_token_434": "997724962668",
444
+ "extra_special_token_435": "558837194455",
445
+ "extra_special_token_436": "163165456396",
446
+ "extra_special_token_437": "378749551722",
447
+ "extra_special_token_438": "161238482259",
448
+ "extra_special_token_439": "754978243758",
449
+ "extra_special_token_440": "195388849133",
450
+ "extra_special_token_441": "229775525672",
451
+ "extra_special_token_442": "262437452884",
452
+ "extra_special_token_443": "441377892146",
453
+ "extra_special_token_444": "451885565366",
454
+ "extra_special_token_445": "981277526855",
455
+ "extra_special_token_446": "762495822823",
456
+ "extra_special_token_447": "368763327262",
457
+ "extra_special_token_448": "757422791351",
458
+ "extra_special_token_449": "636324136426",
459
+ "extra_special_token_450": "214193645583",
460
+ "extra_special_token_451": "412843856172",
461
+ "extra_special_token_452": "179386156569",
462
+ "extra_special_token_453": "756916173536",
463
+ "extra_special_token_454": "892697125149",
464
+ "extra_special_token_455": "625334487352",
465
+ "extra_special_token_456": "941861857715",
466
+ "extra_special_token_457": "887417525236",
467
+ "extra_special_token_458": "649516938598",
468
+ "extra_special_token_459": "717628619782",
469
+ "extra_special_token_460": "438124184139",
470
+ "extra_special_token_461": "547563892268",
471
+ "extra_special_token_462": "856317483891",
472
+ "extra_special_token_463": "313313831273",
473
+ "extra_special_token_464": "371496153876",
474
+ "extra_special_token_465": "587541149322",
475
+ "extra_special_token_466": "265847332563",
476
+ "extra_special_token_467": "449549215429",
477
+ "extra_special_token_468": "163497196769",
478
+ "extra_special_token_469": "861342291298",
479
+ "extra_special_token_470": "268433315926",
480
+ "extra_special_token_471": "774679513717",
481
+ "extra_special_token_472": "851254219729",
482
+ "extra_special_token_473": "583527834464",
483
+ "extra_special_token_474": "488496781997",
484
+ "extra_special_token_475": "556814553861",
485
+ "extra_special_token_476": "482829231639",
486
+ "extra_special_token_477": "618878266619",
487
+ "extra_special_token_478": "147444452794",
488
+ "extra_special_token_479": "949235426629",
489
+ "extra_special_token_480": "357299947518",
490
+ "extra_special_token_481": "175528632226",
491
+ "extra_special_token_482": "645527857972",
492
+ "extra_special_token_483": "186872457894",
493
+ "extra_special_token_484": "552738847828",
494
+ "extra_special_token_485": "626748382482",
495
+ "extra_special_token_486": "921894985642",
496
+ "extra_special_token_487": "943878645871",
497
+ "extra_special_token_488": "859289776479",
498
+ "extra_special_token_489": "614583493135",
499
+ "extra_special_token_490": "933775286797",
500
+ "extra_special_token_491": "332234613346",
501
+ "extra_special_token_492": "325196781219",
502
+ "extra_special_token_493": "142526557681",
503
+ "extra_special_token_494": "356722692178",
504
+ "extra_special_token_495": "449318681694",
505
+ "extra_special_token_496": "687284547244",
506
+ "extra_special_token_497": "947262995132",
507
+ "extra_special_token_498": "893974619684",
508
+ "extra_special_token_499": "797238311233"
509
+ },
510
+ "is_local": false,
511
+ "model_max_length": 1000000000000000019884624838656,
512
+ "pad_token": "<|pad|>",
513
+ "padding_side": "left",
514
+ "tokenizer_class": "GPT2Tokenizer",
515
+ "unk_token": "<|unk|>"
516
+ }
granite-4.0-h-tiny/dp3_attn/train.log ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-03-26 13:28:50,427 [INFO] new_opacus_codex.train_steps: epoch=1 step=5 loss=4.4021
2
+ 2026-03-26 13:29:19,980 [INFO] new_opacus_codex.train_steps: epoch=1 step=10 loss=4.3502
3
+ 2026-03-26 13:29:35,485 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=10 eval_loss=3.3060 duration_sec=15.49
4
+ 2026-03-26 13:30:01,856 [INFO] new_opacus_codex.train_steps: epoch=1 step=15 loss=4.3683
5
+ 2026-03-26 13:30:29,410 [INFO] new_opacus_codex.train_steps: epoch=1 step=20 loss=4.4154
6
+ 2026-03-26 13:30:49,522 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=20 eval_loss=3.2372 duration_sec=20.10
7
+ 2026-03-26 13:31:17,251 [INFO] new_opacus_codex.train_steps: epoch=1 step=25 loss=4.5531
8
+ 2026-03-26 13:31:44,684 [INFO] new_opacus_codex.train_steps: epoch=1 step=30 loss=4.3585
9
+ 2026-03-26 13:32:00,241 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=30 eval_loss=3.0623 duration_sec=15.55
10
+ 2026-03-26 13:32:27,062 [INFO] new_opacus_codex.train_steps: epoch=1 step=35 loss=3.8836
11
+ 2026-03-26 13:32:53,317 [INFO] new_opacus_codex.train_steps: epoch=1 step=40 loss=3.9646
12
+ 2026-03-26 13:33:09,155 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=40 eval_loss=2.7558 duration_sec=15.82
13
+ 2026-03-26 13:33:35,483 [INFO] new_opacus_codex.train_steps: epoch=1 step=45 loss=4.0893
14
+ 2026-03-26 13:34:02,672 [INFO] new_opacus_codex.train_steps: epoch=1 step=50 loss=3.7056
15
+ 2026-03-26 13:34:24,380 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=50 eval_loss=2.4179 duration_sec=21.70
16
+ 2026-03-26 13:34:51,789 [INFO] new_opacus_codex.train_steps: epoch=1 step=55 loss=3.3396
17
+ 2026-03-26 13:35:20,965 [INFO] new_opacus_codex.train_steps: epoch=1 step=60 loss=3.3603
18
+ 2026-03-26 13:35:39,456 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=60 eval_loss=2.1676 duration_sec=18.48
19
+ 2026-03-26 13:36:10,347 [INFO] new_opacus_codex.train_steps: epoch=1 step=65 loss=3.5657
20
+ 2026-03-26 13:36:36,756 [INFO] new_opacus_codex.train_steps: epoch=1 step=70 loss=3.5650
21
+ 2026-03-26 13:36:51,888 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=70 eval_loss=1.9885 duration_sec=15.12
22
+ 2026-03-26 13:37:18,560 [INFO] new_opacus_codex.train_steps: epoch=1 step=75 loss=3.6322
23
+ 2026-03-26 13:37:44,922 [INFO] new_opacus_codex.train_steps: epoch=1 step=80 loss=3.5008
24
+ 2026-03-26 13:38:00,544 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=80 eval_loss=1.8409 duration_sec=15.61
25
+ 2026-03-26 13:38:27,028 [INFO] new_opacus_codex.train_steps: epoch=1 step=85 loss=3.0293
26
+ 2026-03-26 13:38:53,958 [INFO] new_opacus_codex.train_steps: epoch=1 step=90 loss=3.0040
27
+ 2026-03-26 13:39:09,479 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=90 eval_loss=1.7126 duration_sec=15.51
28
+ 2026-03-26 13:39:36,178 [INFO] new_opacus_codex.train_steps: epoch=1 step=95 loss=3.2129
29
+ 2026-03-26 13:40:02,350 [INFO] new_opacus_codex.train_steps: epoch=1 step=100 loss=3.2166
30
+ 2026-03-26 13:40:18,099 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=100 eval_loss=1.5964 duration_sec=15.74
31
+ 2026-03-26 13:40:46,357 [INFO] new_opacus_codex.train_steps: epoch=1 step=105 loss=3.2466
32
+ 2026-03-26 13:41:15,369 [INFO] new_opacus_codex.train_steps: epoch=1 step=110 loss=3.1755
33
+ 2026-03-26 13:41:43,110 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=110 eval_loss=1.4967 duration_sec=27.72
34
+ 2026-03-26 13:42:09,873 [INFO] new_opacus_codex.train_steps: epoch=1 step=115 loss=2.7781
35
+ 2026-03-26 13:42:36,594 [INFO] new_opacus_codex.train_steps: epoch=1 step=120 loss=2.7655
36
+ 2026-03-26 13:42:51,912 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=120 eval_loss=1.4279 duration_sec=15.30
37
+ 2026-03-26 13:43:17,283 [INFO] new_opacus_codex.train_steps: epoch=1 step=125 loss=2.9398
38
+ 2026-03-26 13:43:44,136 [INFO] new_opacus_codex.train_steps: epoch=1 step=130 loss=2.6081
39
+ 2026-03-26 13:43:59,575 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=130 eval_loss=1.3772 duration_sec=15.43
40
+ 2026-03-26 13:44:26,897 [INFO] new_opacus_codex.train_steps: epoch=1 step=135 loss=2.6067
41
+ 2026-03-26 13:44:54,382 [INFO] new_opacus_codex.train_steps: epoch=1 step=140 loss=2.8773
42
+ 2026-03-26 13:45:10,088 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=140 eval_loss=1.3374 duration_sec=15.68
43
+ 2026-03-26 13:45:36,666 [INFO] new_opacus_codex.train_steps: epoch=1 step=145 loss=2.6011
44
+ 2026-03-26 13:46:03,774 [INFO] new_opacus_codex.train_steps: epoch=1 step=150 loss=2.5817
45
+ 2026-03-26 13:46:19,247 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=150 eval_loss=1.3047 duration_sec=15.46
46
+ 2026-03-26 13:46:45,525 [INFO] new_opacus_codex.train_steps: epoch=1 step=155 loss=2.3743
47
+ 2026-03-26 13:47:11,114 [INFO] new_opacus_codex.train_steps: epoch=1 step=160 loss=2.1582
48
+ 2026-03-26 13:47:26,398 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=160 eval_loss=1.2764 duration_sec=15.27
49
+ 2026-03-26 13:47:52,492 [INFO] new_opacus_codex.train_steps: epoch=1 step=165 loss=2.6668
50
+ 2026-03-26 13:48:18,025 [INFO] new_opacus_codex.train_steps: epoch=1 step=170 loss=2.7334
51
+ 2026-03-26 13:48:32,462 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=170 eval_loss=1.2512 duration_sec=14.42
52
+ 2026-03-26 13:48:57,583 [INFO] new_opacus_codex.train_steps: epoch=1 step=175 loss=2.3725
53
+ 2026-03-26 13:49:26,046 [INFO] new_opacus_codex.train_steps: epoch=1 step=180 loss=2.6608
54
+ 2026-03-26 13:49:42,818 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=180 eval_loss=1.2290 duration_sec=16.77
55
+ 2026-03-26 13:50:12,610 [INFO] new_opacus_codex.train_steps: epoch=1 step=185 loss=2.7606
56
+ 2026-03-26 13:50:39,173 [INFO] new_opacus_codex.train_steps: epoch=1 step=190 loss=2.6571
57
+ 2026-03-26 13:50:53,954 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=190 eval_loss=1.2091 duration_sec=14.77
58
+ 2026-03-26 13:51:22,448 [INFO] new_opacus_codex.train_steps: epoch=1 step=195 loss=2.8332
59
+ 2026-03-26 13:51:50,178 [INFO] new_opacus_codex.train_steps: epoch=1 step=200 loss=2.4999
60
+ 2026-03-26 13:52:04,979 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=200 eval_loss=1.1912 duration_sec=14.78
61
+ 2026-03-26 13:52:31,441 [INFO] new_opacus_codex.train_steps: epoch=1 step=205 loss=2.3324
62
+ 2026-03-26 13:52:57,572 [INFO] new_opacus_codex.train_steps: epoch=1 step=210 loss=2.3629
63
+ 2026-03-26 13:53:12,368 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=210 eval_loss=1.1755 duration_sec=14.79
64
+ 2026-03-26 13:53:38,793 [INFO] new_opacus_codex.train_steps: epoch=1 step=215 loss=2.4465
65
+ 2026-03-26 13:54:37,263 [INFO] new_opacus_codex.train_steps: epoch=2 step=220 loss=3.5218
66
+ 2026-03-26 13:54:52,804 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=220 eval_loss=1.1623 duration_sec=15.53
67
+ 2026-03-26 13:55:19,769 [INFO] new_opacus_codex.train_steps: epoch=2 step=225 loss=2.3999
68
+ 2026-03-26 13:55:47,071 [INFO] new_opacus_codex.train_steps: epoch=2 step=230 loss=2.3539
69
+ 2026-03-26 13:56:02,427 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=230 eval_loss=1.1508 duration_sec=15.34
70
+ 2026-03-26 13:56:28,490 [INFO] new_opacus_codex.train_steps: epoch=2 step=235 loss=2.6033
71
+ 2026-03-26 13:56:54,582 [INFO] new_opacus_codex.train_steps: epoch=2 step=240 loss=2.6223
72
+ 2026-03-26 13:57:09,541 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=240 eval_loss=1.1405 duration_sec=14.90
73
+ 2026-03-26 13:57:35,400 [INFO] new_opacus_codex.train_steps: epoch=2 step=245 loss=2.3604
74
+ 2026-03-26 13:58:00,884 [INFO] new_opacus_codex.train_steps: epoch=2 step=250 loss=2.3985
75
+ 2026-03-26 13:58:15,479 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=250 eval_loss=1.1313 duration_sec=14.58
76
+ 2026-03-26 13:58:40,945 [INFO] new_opacus_codex.train_steps: epoch=2 step=255 loss=2.4316
77
+ 2026-03-26 13:59:08,360 [INFO] new_opacus_codex.train_steps: epoch=2 step=260 loss=2.1684
78
+ 2026-03-26 13:59:23,303 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=260 eval_loss=1.1231 duration_sec=14.92
79
+ 2026-03-26 13:59:49,790 [INFO] new_opacus_codex.train_steps: epoch=2 step=265 loss=2.2194
80
+ 2026-03-26 14:00:16,087 [INFO] new_opacus_codex.train_steps: epoch=2 step=270 loss=2.3633
81
+ 2026-03-26 14:00:30,935 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=270 eval_loss=1.1161 duration_sec=14.83
82
+ 2026-03-26 14:00:57,001 [INFO] new_opacus_codex.train_steps: epoch=2 step=275 loss=2.3872
83
+ 2026-03-26 14:01:23,469 [INFO] new_opacus_codex.train_steps: epoch=2 step=280 loss=2.3905
84
+ 2026-03-26 14:01:38,086 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=280 eval_loss=1.1094 duration_sec=14.60
85
+ 2026-03-26 14:02:03,745 [INFO] new_opacus_codex.train_steps: epoch=2 step=285 loss=2.4261
86
+ 2026-03-26 14:02:31,618 [INFO] new_opacus_codex.train_steps: epoch=2 step=290 loss=2.4239
87
+ 2026-03-26 14:02:46,953 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=290 eval_loss=1.1036 duration_sec=15.33
88
+ 2026-03-26 14:03:16,693 [INFO] new_opacus_codex.train_steps: epoch=2 step=295 loss=2.4392
89
+ 2026-03-26 14:03:44,160 [INFO] new_opacus_codex.train_steps: epoch=2 step=300 loss=2.5335
90
+ 2026-03-26 14:04:00,709 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=300 eval_loss=1.0980 duration_sec=16.54
91
+ 2026-03-26 14:04:27,210 [INFO] new_opacus_codex.train_steps: epoch=2 step=305 loss=2.6068
92
+ 2026-03-26 14:04:53,885 [INFO] new_opacus_codex.train_steps: epoch=2 step=310 loss=2.6891
93
+ 2026-03-26 14:05:08,824 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=310 eval_loss=1.0933 duration_sec=14.93
94
+ 2026-03-26 14:05:35,633 [INFO] new_opacus_codex.train_steps: epoch=2 step=315 loss=2.6268
95
+ 2026-03-26 14:06:04,639 [INFO] new_opacus_codex.train_steps: epoch=2 step=320 loss=2.6975
96
+ 2026-03-26 14:06:19,868 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=320 eval_loss=1.0890 duration_sec=15.22
97
+ 2026-03-26 14:06:49,350 [INFO] new_opacus_codex.train_steps: epoch=2 step=325 loss=2.6188
98
+ 2026-03-26 14:07:16,806 [INFO] new_opacus_codex.train_steps: epoch=2 step=330 loss=2.4997
99
+ 2026-03-26 14:07:32,024 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=330 eval_loss=1.0854 duration_sec=15.21
100
+ 2026-03-26 14:07:58,282 [INFO] new_opacus_codex.train_steps: epoch=2 step=335 loss=2.3004
101
+ 2026-03-26 14:08:24,321 [INFO] new_opacus_codex.train_steps: epoch=2 step=340 loss=2.0432
102
+ 2026-03-26 14:08:39,216 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=340 eval_loss=1.0819 duration_sec=14.88
103
+ 2026-03-26 14:09:05,690 [INFO] new_opacus_codex.train_steps: epoch=2 step=345 loss=1.9503
104
+ 2026-03-26 14:09:32,583 [INFO] new_opacus_codex.train_steps: epoch=2 step=350 loss=2.1110
105
+ 2026-03-26 14:09:47,193 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=350 eval_loss=1.0786 duration_sec=14.60
106
+ 2026-03-26 14:10:14,328 [INFO] new_opacus_codex.train_steps: epoch=2 step=355 loss=2.6083
107
+ 2026-03-26 14:10:41,365 [INFO] new_opacus_codex.train_steps: epoch=2 step=360 loss=2.7168
108
+ 2026-03-26 14:10:56,297 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=360 eval_loss=1.0758 duration_sec=14.92
109
+ 2026-03-26 14:11:23,042 [INFO] new_opacus_codex.train_steps: epoch=2 step=365 loss=2.4279
110
+ 2026-03-26 14:11:51,535 [INFO] new_opacus_codex.train_steps: epoch=2 step=370 loss=2.3133
111
+ 2026-03-26 14:12:06,950 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=370 eval_loss=1.0733 duration_sec=15.40
112
+ 2026-03-26 14:12:35,257 [INFO] new_opacus_codex.train_steps: epoch=2 step=375 loss=2.1735
113
+ 2026-03-26 14:13:03,165 [INFO] new_opacus_codex.train_steps: epoch=2 step=380 loss=2.1475
114
+ 2026-03-26 14:13:18,806 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=380 eval_loss=1.0712 duration_sec=15.63
115
+ 2026-03-26 14:13:46,134 [INFO] new_opacus_codex.train_steps: epoch=2 step=385 loss=2.4599
116
+ 2026-03-26 14:14:12,572 [INFO] new_opacus_codex.train_steps: epoch=2 step=390 loss=2.4441
117
+ 2026-03-26 14:14:27,683 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=390 eval_loss=1.0691 duration_sec=15.10
118
+ 2026-03-26 14:14:54,744 [INFO] new_opacus_codex.train_steps: epoch=2 step=395 loss=2.2754
119
+ 2026-03-26 14:15:20,834 [INFO] new_opacus_codex.train_steps: epoch=2 step=400 loss=2.2001
120
+ 2026-03-26 14:15:35,826 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=400 eval_loss=1.0671 duration_sec=14.98
121
+ 2026-03-26 14:16:01,972 [INFO] new_opacus_codex.train_steps: epoch=2 step=405 loss=2.1112
122
+ 2026-03-26 14:16:28,129 [INFO] new_opacus_codex.train_steps: epoch=2 step=410 loss=2.3599
123
+ 2026-03-26 14:16:43,174 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=410 eval_loss=1.0654 duration_sec=15.03
124
+ 2026-03-26 14:17:09,300 [INFO] new_opacus_codex.train_steps: epoch=2 step=415 loss=2.4003
125
+ 2026-03-26 14:17:36,443 [INFO] new_opacus_codex.train_steps: epoch=2 step=420 loss=2.2804
126
+ 2026-03-26 14:17:52,884 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=420 eval_loss=1.0639 duration_sec=16.43
127
+ 2026-03-26 14:18:23,222 [INFO] new_opacus_codex.train_steps: epoch=2 step=425 loss=2.2745
128
+ 2026-03-26 14:18:53,560 [INFO] new_opacus_codex.train_steps: epoch=2 step=430 loss=2.2872
129
+ 2026-03-26 14:19:08,940 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=430 eval_loss=1.0624 duration_sec=15.36
130
+ 2026-03-26 14:19:35,503 [INFO] new_opacus_codex.train_steps: epoch=2 step=435 loss=2.2741
131
+ 2026-03-26 14:20:23,481 [INFO] new_opacus_codex.train_steps: epoch=3 step=440 loss=1.5161
132
+ 2026-03-26 14:20:40,532 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=440 eval_loss=1.0613 duration_sec=17.04
133
+ 2026-03-26 14:21:10,466 [INFO] new_opacus_codex.train_steps: epoch=3 step=445 loss=2.0607
134
+ 2026-03-26 14:21:41,949 [INFO] new_opacus_codex.train_steps: epoch=3 step=450 loss=2.3666
135
+ 2026-03-26 14:21:57,067 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=450 eval_loss=1.0602 duration_sec=15.11
136
+ 2026-03-26 14:22:27,570 [INFO] new_opacus_codex.train_steps: epoch=3 step=455 loss=2.5567
137
+ 2026-03-26 14:22:57,276 [INFO] new_opacus_codex.train_steps: epoch=3 step=460 loss=2.4270
138
+ 2026-03-26 14:23:12,393 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=460 eval_loss=1.0592 duration_sec=15.11
139
+ 2026-03-26 14:23:42,557 [INFO] new_opacus_codex.train_steps: epoch=3 step=465 loss=2.3266
140
+ 2026-03-26 14:24:12,801 [INFO] new_opacus_codex.train_steps: epoch=3 step=470 loss=2.4490
141
+ 2026-03-26 14:24:29,837 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=470 eval_loss=1.0584 duration_sec=17.02
142
+ 2026-03-26 14:25:00,026 [INFO] new_opacus_codex.train_steps: epoch=3 step=475 loss=2.4086
143
+ 2026-03-26 14:25:30,390 [INFO] new_opacus_codex.train_steps: epoch=3 step=480 loss=2.4495
144
+ 2026-03-26 14:25:47,903 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=480 eval_loss=1.0576 duration_sec=17.49
145
+ 2026-03-26 14:26:18,222 [INFO] new_opacus_codex.train_steps: epoch=3 step=485 loss=2.2469
146
+ 2026-03-26 14:26:48,165 [INFO] new_opacus_codex.train_steps: epoch=3 step=490 loss=2.3158
147
+ 2026-03-26 14:27:04,554 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=490 eval_loss=1.0569 duration_sec=16.38
148
+ 2026-03-26 14:27:34,354 [INFO] new_opacus_codex.train_steps: epoch=3 step=495 loss=2.7053
149
+ 2026-03-26 14:28:04,550 [INFO] new_opacus_codex.train_steps: epoch=3 step=500 loss=2.6984
150
+ 2026-03-26 14:28:21,117 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=500 eval_loss=1.0563 duration_sec=16.55
151
+ 2026-03-26 14:28:50,454 [INFO] new_opacus_codex.train_steps: epoch=3 step=505 loss=2.4559
152
+ 2026-03-26 14:29:19,981 [INFO] new_opacus_codex.train_steps: epoch=3 step=510 loss=2.4668
153
+ 2026-03-26 14:29:36,865 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=510 eval_loss=1.0558 duration_sec=16.86
154
+ 2026-03-26 14:30:05,874 [INFO] new_opacus_codex.train_steps: epoch=3 step=515 loss=2.3487
155
+ 2026-03-26 14:30:38,675 [INFO] new_opacus_codex.train_steps: epoch=3 step=520 loss=2.3129
156
+ 2026-03-26 14:30:56,268 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=520 eval_loss=1.0554 duration_sec=17.58
157
+ 2026-03-26 14:31:26,460 [INFO] new_opacus_codex.train_steps: epoch=3 step=525 loss=2.2101
158
+ 2026-03-26 14:31:56,294 [INFO] new_opacus_codex.train_steps: epoch=3 step=530 loss=2.0612
159
+ 2026-03-26 14:32:13,716 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=530 eval_loss=1.0550 duration_sec=17.41
160
+ 2026-03-26 14:32:43,980 [INFO] new_opacus_codex.train_steps: epoch=3 step=535 loss=2.6032
161
+ 2026-03-26 14:33:14,333 [INFO] new_opacus_codex.train_steps: epoch=3 step=540 loss=2.6731
162
+ 2026-03-26 14:33:31,447 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=540 eval_loss=1.0546 duration_sec=17.10
163
+ 2026-03-26 14:34:02,490 [INFO] new_opacus_codex.train_steps: epoch=3 step=545 loss=2.5375
164
+ 2026-03-26 14:34:32,377 [INFO] new_opacus_codex.train_steps: epoch=3 step=550 loss=2.4913
165
+ 2026-03-26 14:34:48,653 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=550 eval_loss=1.0544 duration_sec=16.26
166
+ 2026-03-26 14:35:18,078 [INFO] new_opacus_codex.train_steps: epoch=3 step=555 loss=2.2498
167
+ 2026-03-26 14:35:48,237 [INFO] new_opacus_codex.train_steps: epoch=3 step=560 loss=2.1791
168
+ 2026-03-26 14:36:05,242 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=560 eval_loss=1.0541 duration_sec=16.99
169
+ 2026-03-26 14:36:36,139 [INFO] new_opacus_codex.train_steps: epoch=3 step=565 loss=2.2662
170
+ 2026-03-26 14:37:06,529 [INFO] new_opacus_codex.train_steps: epoch=3 step=570 loss=2.3637
171
+ 2026-03-26 14:37:23,597 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=570 eval_loss=1.0539 duration_sec=17.06
172
+ 2026-03-26 14:37:54,494 [INFO] new_opacus_codex.train_steps: epoch=3 step=575 loss=2.3695
173
+ 2026-03-26 14:38:25,235 [INFO] new_opacus_codex.train_steps: epoch=3 step=580 loss=2.2837
174
+ 2026-03-26 14:38:42,536 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=580 eval_loss=1.0538 duration_sec=17.29
175
+ 2026-03-26 14:39:14,574 [INFO] new_opacus_codex.train_steps: epoch=3 step=585 loss=2.4344
176
+ 2026-03-26 14:39:44,112 [INFO] new_opacus_codex.train_steps: epoch=3 step=590 loss=2.6566
177
+ 2026-03-26 14:40:01,252 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=590 eval_loss=1.0537 duration_sec=17.13
178
+ 2026-03-26 14:40:32,669 [INFO] new_opacus_codex.train_steps: epoch=3 step=595 loss=2.5438
179
+ 2026-03-26 14:41:03,424 [INFO] new_opacus_codex.train_steps: epoch=3 step=600 loss=2.1933
180
+ 2026-03-26 14:41:20,738 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=600 eval_loss=1.0536 duration_sec=17.30
181
+ 2026-03-26 14:41:51,013 [INFO] new_opacus_codex.train_steps: epoch=3 step=605 loss=2.3575
182
+ 2026-03-26 14:42:20,992 [INFO] new_opacus_codex.train_steps: epoch=3 step=610 loss=2.5732
183
+ 2026-03-26 14:42:37,691 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=610 eval_loss=1.0535 duration_sec=16.68
184
+ 2026-03-26 14:43:07,201 [INFO] new_opacus_codex.train_steps: epoch=3 step=615 loss=2.4595
185
+ 2026-03-26 14:43:36,970 [INFO] new_opacus_codex.train_steps: epoch=3 step=620 loss=2.4637
186
+ 2026-03-26 14:43:53,883 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=620 eval_loss=1.0535 duration_sec=16.90
187
+ 2026-03-26 14:44:22,929 [INFO] new_opacus_codex.train_steps: epoch=3 step=625 loss=2.4345
188
+ 2026-03-26 14:44:52,557 [INFO] new_opacus_codex.train_steps: epoch=3 step=630 loss=2.6919
189
+ 2026-03-26 14:45:10,324 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=630 eval_loss=1.0534 duration_sec=17.75
190
+ 2026-03-26 14:45:39,592 [INFO] new_opacus_codex.train_steps: epoch=3 step=635 loss=2.6626
191
+ 2026-03-26 14:46:10,564 [INFO] new_opacus_codex.train_steps: epoch=3 step=640 loss=2.4515
192
+ 2026-03-26 14:46:28,052 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=640 eval_loss=1.0534 duration_sec=17.44
193
+ 2026-03-26 14:46:59,176 [INFO] new_opacus_codex.train_steps: epoch=3 step=645 loss=2.5038
194
+ 2026-03-26 14:47:30,740 [INFO] new_opacus_codex.train_steps: epoch=3 step=650 loss=2.3875
195
+ 2026-03-26 14:47:47,960 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=650 eval_loss=1.0535 duration_sec=17.19
196
+ 2026-03-26 14:48:18,395 [INFO] new_opacus_codex.train_steps: epoch=3 step=655 loss=2.3787
granite-4.0-h-tiny/dp8_attn/adapter/README.md ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ibm-granite/granite-4.0-h-tiny-base
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:ibm-granite/granite-4.0-h-tiny-base
7
+ - lora
8
+ - transformers
9
+ ---
10
+
11
+ # Model Card for Model ID
12
+
13
+ <!-- Provide a quick summary of what the model is/does. -->
14
+
15
+
16
+
17
+ ## Model Details
18
+
19
+ ### Model Description
20
+
21
+ <!-- Provide a longer summary of what this model is. -->
22
+
23
+
24
+
25
+ - **Developed by:** [More Information Needed]
26
+ - **Funded by [optional]:** [More Information Needed]
27
+ - **Shared by [optional]:** [More Information Needed]
28
+ - **Model type:** [More Information Needed]
29
+ - **Language(s) (NLP):** [More Information Needed]
30
+ - **License:** [More Information Needed]
31
+ - **Finetuned from model [optional]:** [More Information Needed]
32
+
33
+ ### Model Sources [optional]
34
+
35
+ <!-- Provide the basic links for the model. -->
36
+
37
+ - **Repository:** [More Information Needed]
38
+ - **Paper [optional]:** [More Information Needed]
39
+ - **Demo [optional]:** [More Information Needed]
40
+
41
+ ## Uses
42
+
43
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
44
+
45
+ ### Direct Use
46
+
47
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
48
+
49
+ [More Information Needed]
50
+
51
+ ### Downstream Use [optional]
52
+
53
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
54
+
55
+ [More Information Needed]
56
+
57
+ ### Out-of-Scope Use
58
+
59
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
60
+
61
+ [More Information Needed]
62
+
63
+ ## Bias, Risks, and Limitations
64
+
65
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
66
+
67
+ [More Information Needed]
68
+
69
+ ### Recommendations
70
+
71
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
72
+
73
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
74
+
75
+ ## How to Get Started with the Model
76
+
77
+ Use the code below to get started with the model.
78
+
79
+ [More Information Needed]
80
+
81
+ ## Training Details
82
+
83
+ ### Training Data
84
+
85
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
86
+
87
+ [More Information Needed]
88
+
89
+ ### Training Procedure
90
+
91
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
92
+
93
+ #### Preprocessing [optional]
94
+
95
+ [More Information Needed]
96
+
97
+
98
+ #### Training Hyperparameters
99
+
100
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
101
+
102
+ #### Speeds, Sizes, Times [optional]
103
+
104
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
105
+
106
+ [More Information Needed]
107
+
108
+ ## Evaluation
109
+
110
+ <!-- This section describes the evaluation protocols and provides the results. -->
111
+
112
+ ### Testing Data, Factors & Metrics
113
+
114
+ #### Testing Data
115
+
116
+ <!-- This should link to a Dataset Card if possible. -->
117
+
118
+ [More Information Needed]
119
+
120
+ #### Factors
121
+
122
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
123
+
124
+ [More Information Needed]
125
+
126
+ #### Metrics
127
+
128
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
129
+
130
+ [More Information Needed]
131
+
132
+ ### Results
133
+
134
+ [More Information Needed]
135
+
136
+ #### Summary
137
+
138
+
139
+
140
+ ## Model Examination [optional]
141
+
142
+ <!-- Relevant interpretability work for the model goes here -->
143
+
144
+ [More Information Needed]
145
+
146
+ ## Environmental Impact
147
+
148
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
149
+
150
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
151
+
152
+ - **Hardware Type:** [More Information Needed]
153
+ - **Hours used:** [More Information Needed]
154
+ - **Cloud Provider:** [More Information Needed]
155
+ - **Compute Region:** [More Information Needed]
156
+ - **Carbon Emitted:** [More Information Needed]
157
+
158
+ ## Technical Specifications [optional]
159
+
160
+ ### Model Architecture and Objective
161
+
162
+ [More Information Needed]
163
+
164
+ ### Compute Infrastructure
165
+
166
+ [More Information Needed]
167
+
168
+ #### Hardware
169
+
170
+ [More Information Needed]
171
+
172
+ #### Software
173
+
174
+ [More Information Needed]
175
+
176
+ ## Citation [optional]
177
+
178
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
179
+
180
+ **BibTeX:**
181
+
182
+ [More Information Needed]
183
+
184
+ **APA:**
185
+
186
+ [More Information Needed]
187
+
188
+ ## Glossary [optional]
189
+
190
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
191
+
192
+ [More Information Needed]
193
+
194
+ ## More Information [optional]
195
+
196
+ [More Information Needed]
197
+
198
+ ## Model Card Authors [optional]
199
+
200
+ [More Information Needed]
201
+
202
+ ## Model Card Contact
203
+
204
+ [More Information Needed]
205
+ ### Framework versions
206
+
207
+ - PEFT 0.18.1
granite-4.0-h-tiny/dp8_attn/adapter/adapter_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "ibm-granite/granite-4.0-h-tiny-base",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": true,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": [
25
+ "lm_head",
26
+ "embed_tokens"
27
+ ],
28
+ "peft_type": "LORA",
29
+ "peft_version": "0.18.1",
30
+ "qalora_group_size": 16,
31
+ "r": 16,
32
+ "rank_pattern": {},
33
+ "revision": null,
34
+ "target_modules": [
35
+ "q_proj",
36
+ "v_proj",
37
+ "o_proj",
38
+ "k_proj"
39
+ ],
40
+ "target_parameters": null,
41
+ "task_type": "CAUSAL_LM",
42
+ "trainable_token_indices": null,
43
+ "use_dora": false,
44
+ "use_qalora": false,
45
+ "use_rslora": false
46
+ }
granite-4.0-h-tiny/dp8_attn/audit_results.json ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "delta": 1e-05,
3
+ "num_canaries": 500,
4
+ "num_members": 250,
5
+ "paper_guess_fraction": 0.2,
6
+ "paper_guess_steps": 20,
7
+ "loss": {
8
+ "auc": 0.508704,
9
+ "empirical_epsilon": {
10
+ "0.05": 0.0,
11
+ "0.01": 0.0
12
+ },
13
+ "empirical_epsilon_details": {
14
+ "0.05": {
15
+ "epsilon": 0.0,
16
+ "num_guesses": 0,
17
+ "correct_guesses": 0,
18
+ "candidate_num_guesses": [
19
+ 5,
20
+ 10,
21
+ 15,
22
+ 20,
23
+ 25,
24
+ 30,
25
+ 35,
26
+ 40,
27
+ 45,
28
+ 50,
29
+ 55,
30
+ 60,
31
+ 65,
32
+ 70,
33
+ 75,
34
+ 80,
35
+ 85,
36
+ 90,
37
+ 95,
38
+ 100
39
+ ],
40
+ "direction": "lower"
41
+ },
42
+ "0.01": {
43
+ "epsilon": 0.0,
44
+ "num_guesses": 0,
45
+ "correct_guesses": 0,
46
+ "candidate_num_guesses": [
47
+ 5,
48
+ 10,
49
+ 15,
50
+ 20,
51
+ 25,
52
+ 30,
53
+ 35,
54
+ 40,
55
+ 45,
56
+ 50,
57
+ 55,
58
+ 60,
59
+ 65,
60
+ 70,
61
+ 75,
62
+ 80,
63
+ 85,
64
+ 90,
65
+ 95,
66
+ 100
67
+ ],
68
+ "direction": "lower"
69
+ }
70
+ }
71
+ },
72
+ "embedding": {
73
+ "auc": 0.521224,
74
+ "empirical_epsilon": {
75
+ "0.05": 0.0,
76
+ "0.01": 0.0
77
+ },
78
+ "empirical_epsilon_details": {
79
+ "0.05": {
80
+ "epsilon": 0.0,
81
+ "num_guesses": 0,
82
+ "correct_guesses": 0,
83
+ "candidate_num_guesses": [
84
+ 5,
85
+ 10,
86
+ 15,
87
+ 20,
88
+ 25,
89
+ 30,
90
+ 35,
91
+ 40,
92
+ 45,
93
+ 50,
94
+ 55,
95
+ 60,
96
+ 65,
97
+ 70,
98
+ 75,
99
+ 80,
100
+ 85,
101
+ 90,
102
+ 95,
103
+ 100
104
+ ],
105
+ "direction": "lower"
106
+ },
107
+ "0.01": {
108
+ "epsilon": 0.0,
109
+ "num_guesses": 0,
110
+ "correct_guesses": 0,
111
+ "candidate_num_guesses": [
112
+ 5,
113
+ 10,
114
+ 15,
115
+ 20,
116
+ 25,
117
+ 30,
118
+ 35,
119
+ 40,
120
+ 45,
121
+ 50,
122
+ 55,
123
+ 60,
124
+ 65,
125
+ 70,
126
+ 75,
127
+ 80,
128
+ 85,
129
+ 90,
130
+ 95,
131
+ 100
132
+ ],
133
+ "direction": "lower"
134
+ }
135
+ }
136
+ }
137
+ }
granite-4.0-h-tiny/dp8_attn/canary_meta.json ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/dp8_attn/metrics.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/dp8_attn/resolved_config.yaml ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ name: ibm-granite/granite-4.0-h-tiny-base
3
+ tokenizer_name: ibm-granite/granite-4.0-h-tiny-base
4
+ max_length: 1024
5
+ dtype: bfloat16
6
+ trust_remote_code: true
7
+ use_fast_tokenizer: true
8
+ cache_dir: null
9
+ local_files_only: false
10
+ low_cpu_mem_usage: true
11
+ tie_word_embeddings: true
12
+ gradient_checkpointing: false
13
+ use_chat_template: false
14
+ dataset:
15
+ name: melihcatal/codedp-cpt
16
+ split: train
17
+ mode: cpt
18
+ text_column: text
19
+ validation_ratio: 0.05
20
+ max_samples: -1
21
+ lora:
22
+ enabled: true
23
+ r: 16
24
+ alpha: 32
25
+ dropout: 0.05
26
+ target_modules:
27
+ - q_proj
28
+ - k_proj
29
+ - v_proj
30
+ - o_proj
31
+ modules_to_save:
32
+ - lm_head
33
+ bias: none
34
+ training:
35
+ seed: 42
36
+ epochs: 3
37
+ warmup_steps: null
38
+ warmup_ratio: 0.05
39
+ mixed_precision: false
40
+ mixed_precision_dtype: bfloat16
41
+ batch_size: 16
42
+ eval_batch_size: 8
43
+ eval_every_steps: 10
44
+ eval_every_epochs: 1
45
+ learning_rate: 0.00017
46
+ optimizer: adamw
47
+ lr_scheduler: cosine
48
+ adam_beta1: 0.9
49
+ adam_beta2: 0.999
50
+ adam_epsilon: 1.0e-08
51
+ sgd_momentum: 0.9
52
+ weight_decay: 0.01
53
+ max_grad_norm: 1.0
54
+ log_every: 5
55
+ gradient_accumulation_steps: 2
56
+ num_workers: 4
57
+ output_dir: /scratch/mcatal/runs/cpt/granite-4.0-h-tiny/dp8_attn
58
+ min_lr_ratio: 0.05
59
+ distributed:
60
+ strategy: dpddp
61
+ backend: nccl
62
+ devices: null
63
+ dp:
64
+ module_validator: auto
65
+ target_delta: 1.0e-05
66
+ noise_multiplier: null
67
+ max_grad_norm: 1.0
68
+ grad_sample_mode: hooks
69
+ clipping: per_layer
70
+ secure_mode: false
71
+ enabled: true
72
+ target_epsilon: 8.0
73
+ audit:
74
+ enabled: true
75
+ run_every_epoch: true
76
+ epoch_device: cuda
77
+ q_canary: auto
78
+ num_canaries: 500
79
+ prefix_length: 49
80
+ num_digits: 12
81
+ batch_size: 32
82
+ delta: 1.0e-05
83
+ p_values:
84
+ - 0.05
85
+ - 0.01
86
+ paper_guess_fraction: 0.2
87
+ paper_guess_steps: 20
88
+ enable_holdout_empirical_epsilon: false
89
+ holdout_seed: 42
90
+ tie_seed: 42
91
+ tracking:
92
+ enabled: true
93
+ tensorboard: true
94
+ wandb: false
95
+ wandb_project: codedp-finetune-h200-audit
96
+ wandb_run_name: granite-4.0-h-tiny-cpt-dp8_attn
97
+ wandb_mode: online
98
+ codecarbon: true
99
+ codecarbon_output_file: codecarbon.csv
100
+ codecarbon_measure_power_secs: 15
101
+ codecarbon_country_iso_code: null
102
+ codecarbon_project_name: codedp-granite-4.0-h-tiny-cpt-dp8_attn
103
+ moe:
104
+ output_router_logits: false
105
+ router_aux_loss_coef: 0.0
106
+ freeze_router: true
107
+ profile:
108
+ enabled: false
109
+ num_batches: 8
110
+ top_experts: 8
111
+ output_file: moe_expert_profile.json
granite-4.0-h-tiny/dp8_attn/summary.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audit/delta": 1e-05,
3
+ "audit/embedding/auc": 0.521224,
4
+ "audit/embedding/empirical_epsilon/0.01": 0.0,
5
+ "audit/embedding/empirical_epsilon/0.05": 0.0,
6
+ "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 0.0,
7
+ "audit/embedding/empirical_epsilon_details/0.01/epsilon": 0.0,
8
+ "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 0.0,
9
+ "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 0.0,
10
+ "audit/embedding/empirical_epsilon_details/0.05/epsilon": 0.0,
11
+ "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 0.0,
12
+ "audit/loss/auc": 0.508704,
13
+ "audit/loss/empirical_epsilon/0.01": 0.0,
14
+ "audit/loss/empirical_epsilon/0.05": 0.0,
15
+ "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 0.0,
16
+ "audit/loss/empirical_epsilon_details/0.01/epsilon": 0.0,
17
+ "audit/loss/empirical_epsilon_details/0.01/num_guesses": 0.0,
18
+ "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 0.0,
19
+ "audit/loss/empirical_epsilon_details/0.05/epsilon": 0.0,
20
+ "audit/loss/empirical_epsilon_details/0.05/num_guesses": 0.0,
21
+ "audit/num_canaries": 500.0,
22
+ "audit/num_members": 250.0,
23
+ "audit/paper_guess_fraction": 0.2,
24
+ "audit/paper_guess_steps": 20.0,
25
+ "energy/codecarbon/cpu_count": 8.0,
26
+ "energy/codecarbon/cpu_energy": 0.09234741829647541,
27
+ "energy/codecarbon/cpu_power": 80.67024766628074,
28
+ "energy/codecarbon/cpu_utilization_percent": 10.131570286788904,
29
+ "energy/codecarbon/duration": 4281.182496272959,
30
+ "energy/codecarbon/emissions": 0.10626002024592303,
31
+ "energy/codecarbon/emissions_rate": 2.4820250091751314e-05,
32
+ "energy/codecarbon/energy_consumed": 3.0496805741733777,
33
+ "energy/codecarbon/gpu_count": 8.0,
34
+ "energy/codecarbon/gpu_energy": 2.9138335321761133,
35
+ "energy/codecarbon/gpu_power": 2456.2229524935556,
36
+ "energy/codecarbon/gpu_utilization_percent": 65.75041137752703,
37
+ "energy/codecarbon/latitude": 47.4843,
38
+ "energy/codecarbon/longitude": 8.212,
39
+ "energy/codecarbon/pue": 1.0,
40
+ "energy/codecarbon/ram_energy": 0.04349962370079205,
41
+ "energy/codecarbon/ram_power": 38.0,
42
+ "energy/codecarbon/ram_total_size": 256.0,
43
+ "energy/codecarbon/ram_used_gb": 528.5185123357518,
44
+ "energy/codecarbon/ram_utilization_percent": 26.66384579219558,
45
+ "energy/codecarbon/water_consumed": 0.0,
46
+ "energy/codecarbon/wue": 0.0,
47
+ "eval/duration_sec": 16.99426259100437,
48
+ "eval/loss": 0.9995726991606795,
49
+ "perf/audit_duration_sec": 7.26710542710498,
50
+ "perf/epoch_duration_sec": 1322.5173796641175,
51
+ "perf/epoch_samples": 61331.0,
52
+ "perf/epoch_samples_per_sec": 46.37443782823963,
53
+ "perf/epoch_tokens": 47090629.0,
54
+ "perf/epoch_tokens_per_sec": 35606.81297962202,
55
+ "perf/logical_batch_size": 35.0,
56
+ "perf/logical_token_count": 28696.0,
57
+ "perf/physical_batches": 3.0,
58
+ "perf/samples_per_sec": 8.308494612014599,
59
+ "perf/step_duration_sec": 4.212556140962988,
60
+ "perf/tokens_per_sec": 6812.016039610598,
61
+ "privacy/epsilon": 7.999857852423159,
62
+ "system/cuda_epoch_peak_memory_gb": 92.61329412460327,
63
+ "system/cuda_max_memory_allocated_gb": 92.61329412460327,
64
+ "system/cuda_memory_allocated_gb": 15.311615943908691,
65
+ "train/epoch_canary_loss": 11.119945305960519,
66
+ "train/epoch_loss": 2.0106413153706493,
67
+ "train/epoch_real_loss": 1.0785744946160816,
68
+ "train/lr": 8.50409354362747e-06,
69
+ "train/step_canary_loss": 10.8125,
70
+ "train/step_loss": 1.8743229457310269,
71
+ "train/step_real_loss": 1.0363688468933105
72
+ }
granite-4.0-h-tiny/dp8_attn/tokenizer/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
granite-4.0-h-tiny/dp8_attn/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,516 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|end_of_text|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|end_of_text|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": {
9
+ "extra_special_token_0": "865331112869",
10
+ "extra_special_token_1": "569765693871",
11
+ "extra_special_token_2": "485177821815",
12
+ "extra_special_token_3": "135441121756",
13
+ "extra_special_token_4": "367459894796",
14
+ "extra_special_token_5": "877482678543",
15
+ "extra_special_token_6": "457919547633",
16
+ "extra_special_token_7": "765474393376",
17
+ "extra_special_token_8": "114848338811",
18
+ "extra_special_token_9": "746285987371",
19
+ "extra_special_token_10": "649291669397",
20
+ "extra_special_token_11": "927914615679",
21
+ "extra_special_token_12": "445925149649",
22
+ "extra_special_token_13": "691587454538",
23
+ "extra_special_token_14": "143777992227",
24
+ "extra_special_token_15": "997981281989",
25
+ "extra_special_token_16": "425949483533",
26
+ "extra_special_token_17": "982993456429",
27
+ "extra_special_token_18": "718726519731",
28
+ "extra_special_token_19": "172599315861",
29
+ "extra_special_token_20": "643489267333",
30
+ "extra_special_token_21": "282322838685",
31
+ "extra_special_token_22": "781653545886",
32
+ "extra_special_token_23": "796415361892",
33
+ "extra_special_token_24": "841991688488",
34
+ "extra_special_token_25": "211411365397",
35
+ "extra_special_token_26": "698218415444",
36
+ "extra_special_token_27": "355977139358",
37
+ "extra_special_token_28": "682564697312",
38
+ "extra_special_token_29": "383837596997",
39
+ "extra_special_token_30": "689362171782",
40
+ "extra_special_token_31": "749966767285",
41
+ "extra_special_token_32": "753159165157",
42
+ "extra_special_token_33": "795693824762",
43
+ "extra_special_token_34": "669689115557",
44
+ "extra_special_token_35": "327491773134",
45
+ "extra_special_token_36": "983569279932",
46
+ "extra_special_token_37": "612128769512",
47
+ "extra_special_token_38": "374327157578",
48
+ "extra_special_token_39": "311632789559",
49
+ "extra_special_token_40": "523918658846",
50
+ "extra_special_token_41": "765981581453",
51
+ "extra_special_token_42": "794825141891",
52
+ "extra_special_token_43": "873898736873",
53
+ "extra_special_token_44": "447445629421",
54
+ "extra_special_token_45": "473822473819",
55
+ "extra_special_token_46": "181439694557",
56
+ "extra_special_token_47": "592538279337",
57
+ "extra_special_token_48": "668134915514",
58
+ "extra_special_token_49": "643692393748",
59
+ "extra_special_token_50": "696651276628",
60
+ "extra_special_token_51": "853859348234",
61
+ "extra_special_token_52": "778466723723",
62
+ "extra_special_token_53": "929826356991",
63
+ "extra_special_token_54": "272362973463",
64
+ "extra_special_token_55": "694235616268",
65
+ "extra_special_token_56": "281673864127",
66
+ "extra_special_token_57": "479676316326",
67
+ "extra_special_token_58": "646979124677",
68
+ "extra_special_token_59": "922327493433",
69
+ "extra_special_token_60": "883685933161",
70
+ "extra_special_token_61": "264259917554",
71
+ "extra_special_token_62": "836746273134",
72
+ "extra_special_token_63": "658481324922",
73
+ "extra_special_token_64": "481884157827",
74
+ "extra_special_token_65": "587787496812",
75
+ "extra_special_token_66": "579184949249",
76
+ "extra_special_token_67": "912193598348",
77
+ "extra_special_token_68": "529679678956",
78
+ "extra_special_token_69": "795838284624",
79
+ "extra_special_token_70": "159337222655",
80
+ "extra_special_token_71": "173781362446",
81
+ "extra_special_token_72": "773687856563",
82
+ "extra_special_token_73": "535787224917",
83
+ "extra_special_token_74": "351885857332",
84
+ "extra_special_token_75": "578827344666",
85
+ "extra_special_token_76": "198462689911",
86
+ "extra_special_token_77": "722618266242",
87
+ "extra_special_token_78": "952872416512",
88
+ "extra_special_token_79": "517778845323",
89
+ "extra_special_token_80": "749665846687",
90
+ "extra_special_token_81": "661436365453",
91
+ "extra_special_token_82": "259666844669",
92
+ "extra_special_token_83": "242851284913",
93
+ "extra_special_token_84": "514532995959",
94
+ "extra_special_token_85": "161588262349",
95
+ "extra_special_token_86": "742765629356",
96
+ "extra_special_token_87": "225164373623",
97
+ "extra_special_token_88": "676539973863",
98
+ "extra_special_token_89": "826214551218",
99
+ "extra_special_token_90": "182345464792",
100
+ "extra_special_token_91": "232776999554",
101
+ "extra_special_token_92": "337326533813",
102
+ "extra_special_token_93": "676676697292",
103
+ "extra_special_token_94": "929185622831",
104
+ "extra_special_token_95": "545512344383",
105
+ "extra_special_token_96": "499444466686",
106
+ "extra_special_token_97": "314697386682",
107
+ "extra_special_token_98": "517379856925",
108
+ "extra_special_token_99": "379557332953",
109
+ "extra_special_token_100": "614797267726",
110
+ "extra_special_token_101": "429781429464",
111
+ "extra_special_token_102": "922466849763",
112
+ "extra_special_token_103": "721737645236",
113
+ "extra_special_token_104": "479227349997",
114
+ "extra_special_token_105": "136931728327",
115
+ "extra_special_token_106": "259533577263",
116
+ "extra_special_token_107": "488538864842",
117
+ "extra_special_token_108": "937495658852",
118
+ "extra_special_token_109": "489991411364",
119
+ "extra_special_token_110": "499148455254",
120
+ "extra_special_token_111": "441373944925",
121
+ "extra_special_token_112": "899151413682",
122
+ "extra_special_token_113": "467893531755",
123
+ "extra_special_token_114": "527117488925",
124
+ "extra_special_token_115": "928335588653",
125
+ "extra_special_token_116": "374439448821",
126
+ "extra_special_token_117": "879425227932",
127
+ "extra_special_token_118": "867678158885",
128
+ "extra_special_token_119": "399749397872",
129
+ "extra_special_token_120": "129693547287",
130
+ "extra_special_token_121": "689285841825",
131
+ "extra_special_token_122": "771619544974",
132
+ "extra_special_token_123": "724883568652",
133
+ "extra_special_token_124": "516968424863",
134
+ "extra_special_token_125": "733737988257",
135
+ "extra_special_token_126": "852347289392",
136
+ "extra_special_token_127": "296953381169",
137
+ "extra_special_token_128": "377273562477",
138
+ "extra_special_token_129": "262296912232",
139
+ "extra_special_token_130": "547149832394",
140
+ "extra_special_token_131": "298464134954",
141
+ "extra_special_token_132": "216667245274",
142
+ "extra_special_token_133": "843998562287",
143
+ "extra_special_token_134": "572154333646",
144
+ "extra_special_token_135": "124589118494",
145
+ "extra_special_token_136": "841824384614",
146
+ "extra_special_token_137": "232896526252",
147
+ "extra_special_token_138": "295448593321",
148
+ "extra_special_token_139": "123741461297",
149
+ "extra_special_token_140": "653573457168",
150
+ "extra_special_token_141": "196735786156",
151
+ "extra_special_token_142": "377338713663",
152
+ "extra_special_token_143": "964342468552",
153
+ "extra_special_token_144": "586855179568",
154
+ "extra_special_token_145": "484773717614",
155
+ "extra_special_token_146": "894885246797",
156
+ "extra_special_token_147": "677896358599",
157
+ "extra_special_token_148": "848845611563",
158
+ "extra_special_token_149": "851852651677",
159
+ "extra_special_token_150": "398549545767",
160
+ "extra_special_token_151": "454244839926",
161
+ "extra_special_token_152": "799364566435",
162
+ "extra_special_token_153": "967114116556",
163
+ "extra_special_token_154": "817378986438",
164
+ "extra_special_token_155": "233795848681",
165
+ "extra_special_token_156": "824387273757",
166
+ "extra_special_token_157": "916198946615",
167
+ "extra_special_token_158": "563117729724",
168
+ "extra_special_token_159": "951794811935",
169
+ "extra_special_token_160": "374598961236",
170
+ "extra_special_token_161": "922867396683",
171
+ "extra_special_token_162": "765737843639",
172
+ "extra_special_token_163": "175469284871",
173
+ "extra_special_token_164": "231853711778",
174
+ "extra_special_token_165": "662426712668",
175
+ "extra_special_token_166": "711412347158",
176
+ "extra_special_token_167": "753466987363",
177
+ "extra_special_token_168": "513361312532",
178
+ "extra_special_token_169": "712992815957",
179
+ "extra_special_token_170": "971621888444",
180
+ "extra_special_token_171": "829235161526",
181
+ "extra_special_token_172": "585544633356",
182
+ "extra_special_token_173": "582471228164",
183
+ "extra_special_token_174": "678666359123",
184
+ "extra_special_token_175": "557533689478",
185
+ "extra_special_token_176": "632962475133",
186
+ "extra_special_token_177": "484489193824",
187
+ "extra_special_token_178": "489562189822",
188
+ "extra_special_token_179": "589547936288",
189
+ "extra_special_token_180": "363214487524",
190
+ "extra_special_token_181": "244885399387",
191
+ "extra_special_token_182": "431751228368",
192
+ "extra_special_token_183": "433581868192",
193
+ "extra_special_token_184": "486391569221",
194
+ "extra_special_token_185": "185438575221",
195
+ "extra_special_token_186": "126574388585",
196
+ "extra_special_token_187": "741757479784",
197
+ "extra_special_token_188": "529854679937",
198
+ "extra_special_token_189": "996116119839",
199
+ "extra_special_token_190": "616248973917",
200
+ "extra_special_token_191": "763531783491",
201
+ "extra_special_token_192": "955456118295",
202
+ "extra_special_token_193": "364196983365",
203
+ "extra_special_token_194": "195792996468",
204
+ "extra_special_token_195": "151859598873",
205
+ "extra_special_token_196": "399223169721",
206
+ "extra_special_token_197": "938488813964",
207
+ "extra_special_token_198": "961981959227",
208
+ "extra_special_token_199": "183368827562",
209
+ "extra_special_token_200": "533417736566",
210
+ "extra_special_token_201": "786391632558",
211
+ "extra_special_token_202": "665661658354",
212
+ "extra_special_token_203": "693281533643",
213
+ "extra_special_token_204": "475794684356",
214
+ "extra_special_token_205": "652154162978",
215
+ "extra_special_token_206": "753233719644",
216
+ "extra_special_token_207": "668514843129",
217
+ "extra_special_token_208": "819162623892",
218
+ "extra_special_token_209": "941169431859",
219
+ "extra_special_token_210": "877385381798",
220
+ "extra_special_token_211": "752644929761",
221
+ "extra_special_token_212": "881136466196",
222
+ "extra_special_token_213": "275597777299",
223
+ "extra_special_token_214": "731681792655",
224
+ "extra_special_token_215": "961133895172",
225
+ "extra_special_token_216": "864718285734",
226
+ "extra_special_token_217": "963852916563",
227
+ "extra_special_token_218": "319584985416",
228
+ "extra_special_token_219": "563365646341",
229
+ "extra_special_token_220": "811371928234",
230
+ "extra_special_token_221": "837131396371",
231
+ "extra_special_token_222": "267514771964",
232
+ "extra_special_token_223": "944513428457",
233
+ "extra_special_token_224": "117298239631",
234
+ "extra_special_token_225": "158142752582",
235
+ "extra_special_token_226": "252867443568",
236
+ "extra_special_token_227": "839269684865",
237
+ "extra_special_token_228": "612788593128",
238
+ "extra_special_token_229": "145669731981",
239
+ "extra_special_token_230": "121557291859",
240
+ "extra_special_token_231": "245416776926",
241
+ "extra_special_token_232": "799417897197",
242
+ "extra_special_token_233": "997958836435",
243
+ "extra_special_token_234": "892336777248",
244
+ "extra_special_token_235": "158929292238",
245
+ "extra_special_token_236": "581976444672",
246
+ "extra_special_token_237": "897784492783",
247
+ "extra_special_token_238": "492373714791",
248
+ "extra_special_token_239": "512659818733",
249
+ "extra_special_token_240": "881112998642",
250
+ "extra_special_token_241": "619454958782",
251
+ "extra_special_token_242": "431149748713",
252
+ "extra_special_token_243": "624221476921",
253
+ "extra_special_token_244": "125866399464",
254
+ "extra_special_token_245": "339882449689",
255
+ "extra_special_token_246": "186198784585",
256
+ "extra_special_token_247": "943193294691",
257
+ "extra_special_token_248": "955668961269",
258
+ "extra_special_token_249": "232787996724",
259
+ "extra_special_token_250": "215671314196",
260
+ "extra_special_token_251": "286173241916",
261
+ "extra_special_token_252": "745977673725",
262
+ "extra_special_token_253": "556976448182",
263
+ "extra_special_token_254": "599961512792",
264
+ "extra_special_token_255": "766294538337",
265
+ "extra_special_token_256": "934912591213",
266
+ "extra_special_token_257": "295118729589",
267
+ "extra_special_token_258": "529455466433",
268
+ "extra_special_token_259": "196119929397",
269
+ "extra_special_token_260": "379571934299",
270
+ "extra_special_token_261": "251789649997",
271
+ "extra_special_token_262": "564544131355",
272
+ "extra_special_token_263": "244371196654",
273
+ "extra_special_token_264": "384598329253",
274
+ "extra_special_token_265": "887753195844",
275
+ "extra_special_token_266": "364947325679",
276
+ "extra_special_token_267": "655517954651",
277
+ "extra_special_token_268": "673948786567",
278
+ "extra_special_token_269": "857231548835",
279
+ "extra_special_token_270": "816115936673",
280
+ "extra_special_token_271": "644234165531",
281
+ "extra_special_token_272": "182782912224",
282
+ "extra_special_token_273": "234316622259",
283
+ "extra_special_token_274": "421369185549",
284
+ "extra_special_token_275": "434632855397",
285
+ "extra_special_token_276": "921889371893",
286
+ "extra_special_token_277": "415956914763",
287
+ "extra_special_token_278": "598916996413",
288
+ "extra_special_token_279": "773671349113",
289
+ "extra_special_token_280": "952465217972",
290
+ "extra_special_token_281": "117657531962",
291
+ "extra_special_token_282": "729825168745",
292
+ "extra_special_token_283": "691315125346",
293
+ "extra_special_token_284": "768461952319",
294
+ "extra_special_token_285": "664847713559",
295
+ "extra_special_token_286": "953267689786",
296
+ "extra_special_token_287": "886464195129",
297
+ "extra_special_token_288": "824488329416",
298
+ "extra_special_token_289": "837873762491",
299
+ "extra_special_token_290": "532833541879",
300
+ "extra_special_token_291": "669183782449",
301
+ "extra_special_token_292": "941976537588",
302
+ "extra_special_token_293": "739394546916",
303
+ "extra_special_token_294": "267954879268",
304
+ "extra_special_token_295": "637551427887",
305
+ "extra_special_token_296": "217756494954",
306
+ "extra_special_token_297": "524444658383",
307
+ "extra_special_token_298": "117783274348",
308
+ "extra_special_token_299": "138218735276",
309
+ "extra_special_token_300": "814611949491",
310
+ "extra_special_token_301": "711641973413",
311
+ "extra_special_token_302": "499156317423",
312
+ "extra_special_token_303": "515856611931",
313
+ "extra_special_token_304": "454164859837",
314
+ "extra_special_token_305": "345271433112",
315
+ "extra_special_token_306": "462294118988",
316
+ "extra_special_token_307": "511785788222",
317
+ "extra_special_token_308": "497294727353",
318
+ "extra_special_token_309": "866519986723",
319
+ "extra_special_token_310": "334513529294",
320
+ "extra_special_token_311": "549946382131",
321
+ "extra_special_token_312": "284445431422",
322
+ "extra_special_token_313": "396521188476",
323
+ "extra_special_token_314": "421435255895",
324
+ "extra_special_token_315": "133373659361",
325
+ "extra_special_token_316": "322683334381",
326
+ "extra_special_token_317": "228358422847",
327
+ "extra_special_token_318": "291762694874",
328
+ "extra_special_token_319": "143182978129",
329
+ "extra_special_token_320": "511923256573",
330
+ "extra_special_token_321": "327158398268",
331
+ "extra_special_token_322": "879764613759",
332
+ "extra_special_token_323": "564395222747",
333
+ "extra_special_token_324": "451161679736",
334
+ "extra_special_token_325": "538631466654",
335
+ "extra_special_token_326": "221762325616",
336
+ "extra_special_token_327": "218391991184",
337
+ "extra_special_token_328": "322589379462",
338
+ "extra_special_token_329": "876537814263",
339
+ "extra_special_token_330": "152676556624",
340
+ "extra_special_token_331": "332522971941",
341
+ "extra_special_token_332": "884354318946",
342
+ "extra_special_token_333": "513349618943",
343
+ "extra_special_token_334": "116639746413",
344
+ "extra_special_token_335": "635185846287",
345
+ "extra_special_token_336": "993832498489",
346
+ "extra_special_token_337": "813981174797",
347
+ "extra_special_token_338": "438745114173",
348
+ "extra_special_token_339": "983493951323",
349
+ "extra_special_token_340": "724492262421",
350
+ "extra_special_token_341": "622553389126",
351
+ "extra_special_token_342": "889965243135",
352
+ "extra_special_token_343": "364492359246",
353
+ "extra_special_token_344": "154962668224",
354
+ "extra_special_token_345": "179564995814",
355
+ "extra_special_token_346": "418412875665",
356
+ "extra_special_token_347": "718951851413",
357
+ "extra_special_token_348": "699446724178",
358
+ "extra_special_token_349": "624266421831",
359
+ "extra_special_token_350": "815458725125",
360
+ "extra_special_token_351": "455423278865",
361
+ "extra_special_token_352": "393741199486",
362
+ "extra_special_token_353": "328552864359",
363
+ "extra_special_token_354": "211662639865",
364
+ "extra_special_token_355": "218784516525",
365
+ "extra_special_token_356": "762486672996",
366
+ "extra_special_token_357": "142799718159",
367
+ "extra_special_token_358": "858146415154",
368
+ "extra_special_token_359": "767858144912",
369
+ "extra_special_token_360": "571317457151",
370
+ "extra_special_token_361": "635127952696",
371
+ "extra_special_token_362": "116427191984",
372
+ "extra_special_token_363": "268921994538",
373
+ "extra_special_token_364": "523937669294",
374
+ "extra_special_token_365": "165429152138",
375
+ "extra_special_token_366": "739246183345",
376
+ "extra_special_token_367": "591464355756",
377
+ "extra_special_token_368": "212985874612",
378
+ "extra_special_token_369": "191887635211",
379
+ "extra_special_token_370": "967214577653",
380
+ "extra_special_token_371": "119342152414",
381
+ "extra_special_token_372": "946444632795",
382
+ "extra_special_token_373": "618423867817",
383
+ "extra_special_token_374": "228565148417",
384
+ "extra_special_token_375": "729116422489",
385
+ "extra_special_token_376": "527874729936",
386
+ "extra_special_token_377": "739784153482",
387
+ "extra_special_token_378": "387763951128",
388
+ "extra_special_token_379": "331369926711",
389
+ "extra_special_token_380": "562716493614",
390
+ "extra_special_token_381": "739667844957",
391
+ "extra_special_token_382": "562389434565",
392
+ "extra_special_token_383": "256497188281",
393
+ "extra_special_token_384": "859927364588",
394
+ "extra_special_token_385": "417668946583",
395
+ "extra_special_token_386": "357621613582",
396
+ "extra_special_token_387": "438435178228",
397
+ "extra_special_token_388": "485692541169",
398
+ "extra_special_token_389": "825815739116",
399
+ "extra_special_token_390": "342221452223",
400
+ "extra_special_token_391": "697747991249",
401
+ "extra_special_token_392": "716763689965",
402
+ "extra_special_token_393": "141499982867",
403
+ "extra_special_token_394": "818479319499",
404
+ "extra_special_token_395": "336813343298",
405
+ "extra_special_token_396": "594688742928",
406
+ "extra_special_token_397": "472129283475",
407
+ "extra_special_token_398": "514354144759",
408
+ "extra_special_token_399": "349249721685",
409
+ "extra_special_token_400": "546276298359",
410
+ "extra_special_token_401": "353755529131",
411
+ "extra_special_token_402": "315534574435",
412
+ "extra_special_token_403": "523723475786",
413
+ "extra_special_token_404": "215826764872",
414
+ "extra_special_token_405": "367968398551",
415
+ "extra_special_token_406": "569853653352",
416
+ "extra_special_token_407": "389715484387",
417
+ "extra_special_token_408": "293847485454",
418
+ "extra_special_token_409": "714738141818",
419
+ "extra_special_token_410": "178478368922",
420
+ "extra_special_token_411": "581493616981",
421
+ "extra_special_token_412": "589439538674",
422
+ "extra_special_token_413": "846657726193",
423
+ "extra_special_token_414": "722339992679",
424
+ "extra_special_token_415": "138154781148",
425
+ "extra_special_token_416": "757785319772",
426
+ "extra_special_token_417": "492516914298",
427
+ "extra_special_token_418": "919181521716",
428
+ "extra_special_token_419": "985781138935",
429
+ "extra_special_token_420": "476969195485",
430
+ "extra_special_token_421": "313145133463",
431
+ "extra_special_token_422": "758963111966",
432
+ "extra_special_token_423": "147541537162",
433
+ "extra_special_token_424": "557163366873",
434
+ "extra_special_token_425": "144373897488",
435
+ "extra_special_token_426": "522515164754",
436
+ "extra_special_token_427": "724964923582",
437
+ "extra_special_token_428": "284776712475",
438
+ "extra_special_token_429": "375429755114",
439
+ "extra_special_token_430": "181233596124",
440
+ "extra_special_token_431": "948585673431",
441
+ "extra_special_token_432": "243165586174",
442
+ "extra_special_token_433": "396847976144",
443
+ "extra_special_token_434": "997724962668",
444
+ "extra_special_token_435": "558837194455",
445
+ "extra_special_token_436": "163165456396",
446
+ "extra_special_token_437": "378749551722",
447
+ "extra_special_token_438": "161238482259",
448
+ "extra_special_token_439": "754978243758",
449
+ "extra_special_token_440": "195388849133",
450
+ "extra_special_token_441": "229775525672",
451
+ "extra_special_token_442": "262437452884",
452
+ "extra_special_token_443": "441377892146",
453
+ "extra_special_token_444": "451885565366",
454
+ "extra_special_token_445": "981277526855",
455
+ "extra_special_token_446": "762495822823",
456
+ "extra_special_token_447": "368763327262",
457
+ "extra_special_token_448": "757422791351",
458
+ "extra_special_token_449": "636324136426",
459
+ "extra_special_token_450": "214193645583",
460
+ "extra_special_token_451": "412843856172",
461
+ "extra_special_token_452": "179386156569",
462
+ "extra_special_token_453": "756916173536",
463
+ "extra_special_token_454": "892697125149",
464
+ "extra_special_token_455": "625334487352",
465
+ "extra_special_token_456": "941861857715",
466
+ "extra_special_token_457": "887417525236",
467
+ "extra_special_token_458": "649516938598",
468
+ "extra_special_token_459": "717628619782",
469
+ "extra_special_token_460": "438124184139",
470
+ "extra_special_token_461": "547563892268",
471
+ "extra_special_token_462": "856317483891",
472
+ "extra_special_token_463": "313313831273",
473
+ "extra_special_token_464": "371496153876",
474
+ "extra_special_token_465": "587541149322",
475
+ "extra_special_token_466": "265847332563",
476
+ "extra_special_token_467": "449549215429",
477
+ "extra_special_token_468": "163497196769",
478
+ "extra_special_token_469": "861342291298",
479
+ "extra_special_token_470": "268433315926",
480
+ "extra_special_token_471": "774679513717",
481
+ "extra_special_token_472": "851254219729",
482
+ "extra_special_token_473": "583527834464",
483
+ "extra_special_token_474": "488496781997",
484
+ "extra_special_token_475": "556814553861",
485
+ "extra_special_token_476": "482829231639",
486
+ "extra_special_token_477": "618878266619",
487
+ "extra_special_token_478": "147444452794",
488
+ "extra_special_token_479": "949235426629",
489
+ "extra_special_token_480": "357299947518",
490
+ "extra_special_token_481": "175528632226",
491
+ "extra_special_token_482": "645527857972",
492
+ "extra_special_token_483": "186872457894",
493
+ "extra_special_token_484": "552738847828",
494
+ "extra_special_token_485": "626748382482",
495
+ "extra_special_token_486": "921894985642",
496
+ "extra_special_token_487": "943878645871",
497
+ "extra_special_token_488": "859289776479",
498
+ "extra_special_token_489": "614583493135",
499
+ "extra_special_token_490": "933775286797",
500
+ "extra_special_token_491": "332234613346",
501
+ "extra_special_token_492": "325196781219",
502
+ "extra_special_token_493": "142526557681",
503
+ "extra_special_token_494": "356722692178",
504
+ "extra_special_token_495": "449318681694",
505
+ "extra_special_token_496": "687284547244",
506
+ "extra_special_token_497": "947262995132",
507
+ "extra_special_token_498": "893974619684",
508
+ "extra_special_token_499": "797238311233"
509
+ },
510
+ "is_local": false,
511
+ "model_max_length": 1000000000000000019884624838656,
512
+ "pad_token": "<|pad|>",
513
+ "padding_side": "left",
514
+ "tokenizer_class": "GPT2Tokenizer",
515
+ "unk_token": "<|unk|>"
516
+ }
granite-4.0-h-tiny/dp8_attn/train.log ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-03-26 17:26:21,623 [INFO] new_opacus_codex.train_steps: epoch=1 step=5 loss=4.4037
2
+ 2026-03-26 17:26:48,735 [INFO] new_opacus_codex.train_steps: epoch=1 step=10 loss=4.3490
3
+ 2026-03-26 17:27:06,138 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=10 eval_loss=3.3031 duration_sec=17.39
4
+ 2026-03-26 17:27:30,458 [INFO] new_opacus_codex.train_steps: epoch=1 step=15 loss=4.3648
5
+ 2026-03-26 17:27:54,061 [INFO] new_opacus_codex.train_steps: epoch=1 step=20 loss=4.4076
6
+ 2026-03-26 17:28:10,905 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=20 eval_loss=3.2204 duration_sec=16.84
7
+ 2026-03-26 17:28:33,671 [INFO] new_opacus_codex.train_steps: epoch=1 step=25 loss=4.5385
8
+ 2026-03-26 17:28:56,794 [INFO] new_opacus_codex.train_steps: epoch=1 step=30 loss=4.3479
9
+ 2026-03-26 17:29:13,954 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=30 eval_loss=3.0424 duration_sec=17.15
10
+ 2026-03-26 17:29:36,034 [INFO] new_opacus_codex.train_steps: epoch=1 step=35 loss=3.8731
11
+ 2026-03-26 17:29:58,596 [INFO] new_opacus_codex.train_steps: epoch=1 step=40 loss=3.9419
12
+ 2026-03-26 17:30:15,362 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=40 eval_loss=2.7359 duration_sec=16.75
13
+ 2026-03-26 17:30:37,454 [INFO] new_opacus_codex.train_steps: epoch=1 step=45 loss=4.0604
14
+ 2026-03-26 17:31:00,149 [INFO] new_opacus_codex.train_steps: epoch=1 step=50 loss=3.6684
15
+ 2026-03-26 17:31:17,674 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=50 eval_loss=2.3742 duration_sec=17.51
16
+ 2026-03-26 17:31:39,753 [INFO] new_opacus_codex.train_steps: epoch=1 step=55 loss=3.2821
17
+ 2026-03-26 17:32:01,629 [INFO] new_opacus_codex.train_steps: epoch=1 step=60 loss=3.2626
18
+ 2026-03-26 17:32:18,497 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=60 eval_loss=1.9961 duration_sec=16.86
19
+ 2026-03-26 17:32:41,426 [INFO] new_opacus_codex.train_steps: epoch=1 step=65 loss=3.4029
20
+ 2026-03-26 17:33:04,379 [INFO] new_opacus_codex.train_steps: epoch=1 step=70 loss=3.3352
21
+ 2026-03-26 17:33:22,252 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=70 eval_loss=1.6862 duration_sec=17.86
22
+ 2026-03-26 17:33:44,837 [INFO] new_opacus_codex.train_steps: epoch=1 step=75 loss=3.3517
23
+ 2026-03-26 17:34:07,591 [INFO] new_opacus_codex.train_steps: epoch=1 step=80 loss=3.1838
24
+ 2026-03-26 17:34:24,885 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=80 eval_loss=1.4886 duration_sec=17.28
25
+ 2026-03-26 17:34:46,968 [INFO] new_opacus_codex.train_steps: epoch=1 step=85 loss=2.6801
26
+ 2026-03-26 17:35:09,434 [INFO] new_opacus_codex.train_steps: epoch=1 step=90 loss=2.6484
27
+ 2026-03-26 17:35:26,566 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=90 eval_loss=1.3736 duration_sec=17.10
28
+ 2026-03-26 17:35:48,874 [INFO] new_opacus_codex.train_steps: epoch=1 step=95 loss=2.8568
29
+ 2026-03-26 17:36:13,765 [INFO] new_opacus_codex.train_steps: epoch=1 step=100 loss=2.8620
30
+ 2026-03-26 17:36:30,768 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=100 eval_loss=1.2965 duration_sec=16.99
31
+ 2026-03-26 17:36:53,427 [INFO] new_opacus_codex.train_steps: epoch=1 step=105 loss=2.9108
32
+ 2026-03-26 17:37:18,876 [INFO] new_opacus_codex.train_steps: epoch=1 step=110 loss=2.8497
33
+ 2026-03-26 17:37:35,771 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=110 eval_loss=1.2415 duration_sec=16.88
34
+ 2026-03-26 17:37:58,033 [INFO] new_opacus_codex.train_steps: epoch=1 step=115 loss=2.4601
35
+ 2026-03-26 17:38:19,907 [INFO] new_opacus_codex.train_steps: epoch=1 step=120 loss=2.4553
36
+ 2026-03-26 17:38:37,297 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=120 eval_loss=1.1993 duration_sec=17.37
37
+ 2026-03-26 17:38:59,359 [INFO] new_opacus_codex.train_steps: epoch=1 step=125 loss=2.6337
38
+ 2026-03-26 17:39:20,909 [INFO] new_opacus_codex.train_steps: epoch=1 step=130 loss=2.3083
39
+ 2026-03-26 17:39:38,154 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=130 eval_loss=1.1650 duration_sec=17.23
40
+ 2026-03-26 17:40:00,101 [INFO] new_opacus_codex.train_steps: epoch=1 step=135 loss=2.3016
41
+ 2026-03-26 17:40:21,850 [INFO] new_opacus_codex.train_steps: epoch=1 step=140 loss=2.5665
42
+ 2026-03-26 17:40:38,489 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=140 eval_loss=1.1389 duration_sec=16.62
43
+ 2026-03-26 17:40:59,854 [INFO] new_opacus_codex.train_steps: epoch=1 step=145 loss=2.3101
44
+ 2026-03-26 17:41:21,779 [INFO] new_opacus_codex.train_steps: epoch=1 step=150 loss=2.2803
45
+ 2026-03-26 17:41:39,138 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=150 eval_loss=1.1191 duration_sec=17.35
46
+ 2026-03-26 17:42:00,543 [INFO] new_opacus_codex.train_steps: epoch=1 step=155 loss=2.0894
47
+ 2026-03-26 17:42:22,164 [INFO] new_opacus_codex.train_steps: epoch=1 step=160 loss=1.9007
48
+ 2026-03-26 17:42:39,134 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=160 eval_loss=1.1035 duration_sec=16.95
49
+ 2026-03-26 17:43:01,309 [INFO] new_opacus_codex.train_steps: epoch=1 step=165 loss=2.3738
50
+ 2026-03-26 17:43:23,325 [INFO] new_opacus_codex.train_steps: epoch=1 step=170 loss=2.4276
51
+ 2026-03-26 17:43:41,113 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=170 eval_loss=1.0907 duration_sec=17.77
52
+ 2026-03-26 17:44:02,913 [INFO] new_opacus_codex.train_steps: epoch=1 step=175 loss=2.0950
53
+ 2026-03-26 17:44:25,255 [INFO] new_opacus_codex.train_steps: epoch=1 step=180 loss=2.3697
54
+ 2026-03-26 17:44:41,546 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=180 eval_loss=1.0798 duration_sec=16.28
55
+ 2026-03-26 17:45:03,278 [INFO] new_opacus_codex.train_steps: epoch=1 step=185 loss=2.4739
56
+ 2026-03-26 17:45:25,316 [INFO] new_opacus_codex.train_steps: epoch=1 step=190 loss=2.3745
57
+ 2026-03-26 17:45:41,950 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=190 eval_loss=1.0708 duration_sec=16.62
58
+ 2026-03-26 17:46:03,552 [INFO] new_opacus_codex.train_steps: epoch=1 step=195 loss=2.5070
59
+ 2026-03-26 17:46:25,748 [INFO] new_opacus_codex.train_steps: epoch=1 step=200 loss=2.2084
60
+ 2026-03-26 17:46:42,099 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=200 eval_loss=1.0628 duration_sec=16.34
61
+ 2026-03-26 17:47:04,788 [INFO] new_opacus_codex.train_steps: epoch=1 step=205 loss=2.0751
62
+ 2026-03-26 17:47:26,320 [INFO] new_opacus_codex.train_steps: epoch=1 step=210 loss=2.0969
63
+ 2026-03-26 17:47:42,593 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=210 eval_loss=1.0561 duration_sec=16.26
64
+ 2026-03-26 17:48:04,351 [INFO] new_opacus_codex.train_steps: epoch=1 step=215 loss=2.1621
65
+ 2026-03-26 17:48:49,233 [INFO] new_opacus_codex.train_steps: epoch=2 step=220 loss=3.0754
66
+ 2026-03-26 17:49:05,580 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=220 eval_loss=1.0502 duration_sec=16.34
67
+ 2026-03-26 17:49:27,001 [INFO] new_opacus_codex.train_steps: epoch=2 step=225 loss=2.1009
68
+ 2026-03-26 17:49:48,762 [INFO] new_opacus_codex.train_steps: epoch=2 step=230 loss=2.0643
69
+ 2026-03-26 17:50:05,226 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=230 eval_loss=1.0454 duration_sec=16.45
70
+ 2026-03-26 17:50:27,007 [INFO] new_opacus_codex.train_steps: epoch=2 step=235 loss=2.2919
71
+ 2026-03-26 17:50:49,126 [INFO] new_opacus_codex.train_steps: epoch=2 step=240 loss=2.3218
72
+ 2026-03-26 17:51:05,694 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=240 eval_loss=1.0410 duration_sec=16.56
73
+ 2026-03-26 17:51:27,687 [INFO] new_opacus_codex.train_steps: epoch=2 step=245 loss=2.0916
74
+ 2026-03-26 17:51:48,987 [INFO] new_opacus_codex.train_steps: epoch=2 step=250 loss=2.1139
75
+ 2026-03-26 17:52:05,939 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=250 eval_loss=1.0374 duration_sec=16.94
76
+ 2026-03-26 17:52:27,519 [INFO] new_opacus_codex.train_steps: epoch=2 step=255 loss=2.1325
77
+ 2026-03-26 17:52:48,397 [INFO] new_opacus_codex.train_steps: epoch=2 step=260 loss=1.8994
78
+ 2026-03-26 17:53:05,360 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=260 eval_loss=1.0340 duration_sec=16.95
79
+ 2026-03-26 17:53:26,541 [INFO] new_opacus_codex.train_steps: epoch=2 step=265 loss=1.9436
80
+ 2026-03-26 17:53:48,575 [INFO] new_opacus_codex.train_steps: epoch=2 step=270 loss=2.0555
81
+ 2026-03-26 17:54:05,541 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=270 eval_loss=1.0307 duration_sec=16.95
82
+ 2026-03-26 17:54:27,135 [INFO] new_opacus_codex.train_steps: epoch=2 step=275 loss=2.0832
83
+ 2026-03-26 17:54:48,565 [INFO] new_opacus_codex.train_steps: epoch=2 step=280 loss=2.0941
84
+ 2026-03-26 17:55:04,951 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=280 eval_loss=1.0277 duration_sec=16.37
85
+ 2026-03-26 17:55:26,654 [INFO] new_opacus_codex.train_steps: epoch=2 step=285 loss=2.1160
86
+ 2026-03-26 17:55:48,098 [INFO] new_opacus_codex.train_steps: epoch=2 step=290 loss=2.1182
87
+ 2026-03-26 17:56:05,307 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=290 eval_loss=1.0251 duration_sec=17.19
88
+ 2026-03-26 17:56:27,046 [INFO] new_opacus_codex.train_steps: epoch=2 step=295 loss=2.1196
89
+ 2026-03-26 17:56:48,820 [INFO] new_opacus_codex.train_steps: epoch=2 step=300 loss=2.1963
90
+ 2026-03-26 17:57:05,292 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=300 eval_loss=1.0228 duration_sec=16.46
91
+ 2026-03-26 17:57:26,963 [INFO] new_opacus_codex.train_steps: epoch=2 step=305 loss=2.2534
92
+ 2026-03-26 17:57:49,894 [INFO] new_opacus_codex.train_steps: epoch=2 step=310 loss=2.3017
93
+ 2026-03-26 17:58:06,031 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=310 eval_loss=1.0209 duration_sec=16.13
94
+ 2026-03-26 17:58:29,266 [INFO] new_opacus_codex.train_steps: epoch=2 step=315 loss=2.2543
95
+ 2026-03-26 17:58:51,432 [INFO] new_opacus_codex.train_steps: epoch=2 step=320 loss=2.2957
96
+ 2026-03-26 17:59:08,333 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=320 eval_loss=1.0190 duration_sec=16.89
97
+ 2026-03-26 17:59:30,278 [INFO] new_opacus_codex.train_steps: epoch=2 step=325 loss=2.2357
98
+ 2026-03-26 17:59:53,771 [INFO] new_opacus_codex.train_steps: epoch=2 step=330 loss=2.1549
99
+ 2026-03-26 18:00:10,849 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=330 eval_loss=1.0172 duration_sec=17.07
100
+ 2026-03-26 18:00:33,333 [INFO] new_opacus_codex.train_steps: epoch=2 step=335 loss=1.9978
101
+ 2026-03-26 18:00:55,065 [INFO] new_opacus_codex.train_steps: epoch=2 step=340 loss=1.7871
102
+ 2026-03-26 18:01:12,167 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=340 eval_loss=1.0156 duration_sec=17.09
103
+ 2026-03-26 18:01:33,509 [INFO] new_opacus_codex.train_steps: epoch=2 step=345 loss=1.6894
104
+ 2026-03-26 18:01:55,159 [INFO] new_opacus_codex.train_steps: epoch=2 step=350 loss=1.8369
105
+ 2026-03-26 18:02:11,776 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=350 eval_loss=1.0141 duration_sec=16.61
106
+ 2026-03-26 18:02:33,356 [INFO] new_opacus_codex.train_steps: epoch=2 step=355 loss=2.2433
107
+ 2026-03-26 18:02:55,102 [INFO] new_opacus_codex.train_steps: epoch=2 step=360 loss=2.2975
108
+ 2026-03-26 18:03:12,264 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=360 eval_loss=1.0125 duration_sec=17.15
109
+ 2026-03-26 18:03:33,864 [INFO] new_opacus_codex.train_steps: epoch=2 step=365 loss=2.0665
110
+ 2026-03-26 18:03:55,587 [INFO] new_opacus_codex.train_steps: epoch=2 step=370 loss=1.9807
111
+ 2026-03-26 18:04:13,582 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=370 eval_loss=1.0113 duration_sec=17.99
112
+ 2026-03-26 18:04:35,137 [INFO] new_opacus_codex.train_steps: epoch=2 step=375 loss=1.8653
113
+ 2026-03-26 18:04:57,470 [INFO] new_opacus_codex.train_steps: epoch=2 step=380 loss=1.8529
114
+ 2026-03-26 18:05:14,164 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=380 eval_loss=1.0102 duration_sec=16.68
115
+ 2026-03-26 18:05:35,660 [INFO] new_opacus_codex.train_steps: epoch=2 step=385 loss=2.1010
116
+ 2026-03-26 18:05:57,408 [INFO] new_opacus_codex.train_steps: epoch=2 step=390 loss=2.0590
117
+ 2026-03-26 18:06:13,895 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=390 eval_loss=1.0091 duration_sec=16.47
118
+ 2026-03-26 18:06:35,421 [INFO] new_opacus_codex.train_steps: epoch=2 step=395 loss=1.9242
119
+ 2026-03-26 18:06:58,676 [INFO] new_opacus_codex.train_steps: epoch=2 step=400 loss=1.8661
120
+ 2026-03-26 18:07:15,374 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=400 eval_loss=1.0080 duration_sec=16.68
121
+ 2026-03-26 18:07:36,492 [INFO] new_opacus_codex.train_steps: epoch=2 step=405 loss=1.8016
122
+ 2026-03-26 18:07:58,586 [INFO] new_opacus_codex.train_steps: epoch=2 step=410 loss=1.9926
123
+ 2026-03-26 18:08:16,299 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=410 eval_loss=1.0071 duration_sec=17.69
124
+ 2026-03-26 18:08:37,640 [INFO] new_opacus_codex.train_steps: epoch=2 step=415 loss=2.0021
125
+ 2026-03-26 18:08:59,478 [INFO] new_opacus_codex.train_steps: epoch=2 step=420 loss=1.9054
126
+ 2026-03-26 18:09:16,893 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=420 eval_loss=1.0063 duration_sec=17.40
127
+ 2026-03-26 18:09:38,189 [INFO] new_opacus_codex.train_steps: epoch=2 step=425 loss=1.9168
128
+ 2026-03-26 18:09:59,617 [INFO] new_opacus_codex.train_steps: epoch=2 step=430 loss=1.9480
129
+ 2026-03-26 18:10:16,219 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=430 eval_loss=1.0056 duration_sec=16.58
130
+ 2026-03-26 18:10:37,146 [INFO] new_opacus_codex.train_steps: epoch=2 step=435 loss=1.9287
131
+ 2026-03-26 18:11:22,354 [INFO] new_opacus_codex.train_steps: epoch=3 step=440 loss=1.3197
132
+ 2026-03-26 18:11:38,510 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=440 eval_loss=1.0050 duration_sec=16.14
133
+ 2026-03-26 18:12:00,048 [INFO] new_opacus_codex.train_steps: epoch=3 step=445 loss=1.7475
134
+ 2026-03-26 18:12:21,337 [INFO] new_opacus_codex.train_steps: epoch=3 step=450 loss=1.9872
135
+ 2026-03-26 18:12:38,072 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=450 eval_loss=1.0043 duration_sec=16.73
136
+ 2026-03-26 18:12:59,771 [INFO] new_opacus_codex.train_steps: epoch=3 step=455 loss=2.1328
137
+ 2026-03-26 18:13:20,789 [INFO] new_opacus_codex.train_steps: epoch=3 step=460 loss=2.0179
138
+ 2026-03-26 18:13:37,724 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=460 eval_loss=1.0038 duration_sec=16.92
139
+ 2026-03-26 18:13:59,215 [INFO] new_opacus_codex.train_steps: epoch=3 step=465 loss=1.9457
140
+ 2026-03-26 18:14:20,504 [INFO] new_opacus_codex.train_steps: epoch=3 step=470 loss=2.0557
141
+ 2026-03-26 18:14:36,733 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=470 eval_loss=1.0032 duration_sec=16.21
142
+ 2026-03-26 18:14:58,511 [INFO] new_opacus_codex.train_steps: epoch=3 step=475 loss=2.0081
143
+ 2026-03-26 18:15:20,262 [INFO] new_opacus_codex.train_steps: epoch=3 step=480 loss=2.0313
144
+ 2026-03-26 18:15:37,471 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=480 eval_loss=1.0028 duration_sec=17.20
145
+ 2026-03-26 18:15:59,304 [INFO] new_opacus_codex.train_steps: epoch=3 step=485 loss=1.8972
146
+ 2026-03-26 18:16:21,144 [INFO] new_opacus_codex.train_steps: epoch=3 step=490 loss=1.9504
147
+ 2026-03-26 18:16:38,415 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=490 eval_loss=1.0024 duration_sec=17.26
148
+ 2026-03-26 18:17:00,458 [INFO] new_opacus_codex.train_steps: epoch=3 step=495 loss=2.2316
149
+ 2026-03-26 18:17:22,261 [INFO] new_opacus_codex.train_steps: epoch=3 step=500 loss=2.2176
150
+ 2026-03-26 18:17:39,936 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=500 eval_loss=1.0021 duration_sec=17.66
151
+ 2026-03-26 18:18:01,718 [INFO] new_opacus_codex.train_steps: epoch=3 step=505 loss=2.0301
152
+ 2026-03-26 18:18:23,307 [INFO] new_opacus_codex.train_steps: epoch=3 step=510 loss=2.0475
153
+ 2026-03-26 18:18:40,339 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=510 eval_loss=1.0018 duration_sec=17.02
154
+ 2026-03-26 18:19:01,746 [INFO] new_opacus_codex.train_steps: epoch=3 step=515 loss=1.9562
155
+ 2026-03-26 18:19:23,801 [INFO] new_opacus_codex.train_steps: epoch=3 step=520 loss=1.9296
156
+ 2026-03-26 18:19:40,579 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=520 eval_loss=1.0016 duration_sec=16.69
157
+ 2026-03-26 18:20:01,901 [INFO] new_opacus_codex.train_steps: epoch=3 step=525 loss=1.8632
158
+ 2026-03-26 18:20:24,831 [INFO] new_opacus_codex.train_steps: epoch=3 step=530 loss=1.7505
159
+ 2026-03-26 18:20:41,175 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=530 eval_loss=1.0012 duration_sec=16.33
160
+ 2026-03-26 18:21:02,428 [INFO] new_opacus_codex.train_steps: epoch=3 step=535 loss=2.1539
161
+ 2026-03-26 18:21:24,738 [INFO] new_opacus_codex.train_steps: epoch=3 step=540 loss=2.2057
162
+ 2026-03-26 18:21:41,570 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=540 eval_loss=1.0010 duration_sec=16.82
163
+ 2026-03-26 18:22:03,978 [INFO] new_opacus_codex.train_steps: epoch=3 step=545 loss=2.0904
164
+ 2026-03-26 18:22:27,569 [INFO] new_opacus_codex.train_steps: epoch=3 step=550 loss=2.0540
165
+ 2026-03-26 18:22:44,616 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=550 eval_loss=1.0008 duration_sec=17.04
166
+ 2026-03-26 18:23:06,177 [INFO] new_opacus_codex.train_steps: epoch=3 step=555 loss=1.8775
167
+ 2026-03-26 18:23:28,261 [INFO] new_opacus_codex.train_steps: epoch=3 step=560 loss=1.8288
168
+ 2026-03-26 18:23:45,480 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=560 eval_loss=1.0006 duration_sec=17.20
169
+ 2026-03-26 18:24:07,451 [INFO] new_opacus_codex.train_steps: epoch=3 step=565 loss=1.9104
170
+ 2026-03-26 18:24:28,844 [INFO] new_opacus_codex.train_steps: epoch=3 step=570 loss=1.9790
171
+ 2026-03-26 18:24:45,156 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=570 eval_loss=1.0005 duration_sec=16.30
172
+ 2026-03-26 18:25:07,126 [INFO] new_opacus_codex.train_steps: epoch=3 step=575 loss=1.9842
173
+ 2026-03-26 18:25:28,788 [INFO] new_opacus_codex.train_steps: epoch=3 step=580 loss=1.9075
174
+ 2026-03-26 18:25:45,450 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=580 eval_loss=1.0003 duration_sec=16.65
175
+ 2026-03-26 18:26:07,405 [INFO] new_opacus_codex.train_steps: epoch=3 step=585 loss=2.0189
176
+ 2026-03-26 18:26:29,234 [INFO] new_opacus_codex.train_steps: epoch=3 step=590 loss=2.1962
177
+ 2026-03-26 18:26:46,342 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=590 eval_loss=1.0001 duration_sec=17.09
178
+ 2026-03-26 18:27:08,435 [INFO] new_opacus_codex.train_steps: epoch=3 step=595 loss=2.0936
179
+ 2026-03-26 18:27:30,175 [INFO] new_opacus_codex.train_steps: epoch=3 step=600 loss=1.8410
180
+ 2026-03-26 18:27:47,709 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=600 eval_loss=1.0000 duration_sec=17.53
181
+ 2026-03-26 18:28:09,920 [INFO] new_opacus_codex.train_steps: epoch=3 step=605 loss=1.9650
182
+ 2026-03-26 18:28:31,833 [INFO] new_opacus_codex.train_steps: epoch=3 step=610 loss=2.1060
183
+ 2026-03-26 18:28:48,969 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=610 eval_loss=1.0000 duration_sec=17.13
184
+ 2026-03-26 18:29:10,653 [INFO] new_opacus_codex.train_steps: epoch=3 step=615 loss=2.0100
185
+ 2026-03-26 18:29:31,939 [INFO] new_opacus_codex.train_steps: epoch=3 step=620 loss=2.0135
186
+ 2026-03-26 18:29:48,519 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=620 eval_loss=0.9998 duration_sec=16.57
187
+ 2026-03-26 18:30:10,105 [INFO] new_opacus_codex.train_steps: epoch=3 step=625 loss=2.0115
188
+ 2026-03-26 18:30:31,531 [INFO] new_opacus_codex.train_steps: epoch=3 step=630 loss=2.1898
189
+ 2026-03-26 18:30:48,482 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=630 eval_loss=0.9998 duration_sec=16.94
190
+ 2026-03-26 18:31:09,613 [INFO] new_opacus_codex.train_steps: epoch=3 step=635 loss=2.1663
191
+ 2026-03-26 18:31:31,201 [INFO] new_opacus_codex.train_steps: epoch=3 step=640 loss=2.0312
192
+ 2026-03-26 18:31:47,436 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=640 eval_loss=0.9996 duration_sec=16.22
193
+ 2026-03-26 18:32:09,414 [INFO] new_opacus_codex.train_steps: epoch=3 step=645 loss=2.0599
194
+ 2026-03-26 18:32:31,040 [INFO] new_opacus_codex.train_steps: epoch=3 step=650 loss=1.9682
195
+ 2026-03-26 18:32:47,820 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=650 eval_loss=0.9996 duration_sec=16.77
196
+ 2026-03-26 18:33:09,495 [INFO] new_opacus_codex.train_steps: epoch=3 step=655 loss=1.9618
llama3.1-8b/dp3/adapter/README.md ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.1-8B
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:meta-llama/Llama-3.1-8B
7
+ - lora
8
+ - transformers
9
+ ---
10
+
11
+ # Model Card for Model ID
12
+
13
+ <!-- Provide a quick summary of what the model is/does. -->
14
+
15
+
16
+
17
+ ## Model Details
18
+
19
+ ### Model Description
20
+
21
+ <!-- Provide a longer summary of what this model is. -->
22
+
23
+
24
+
25
+ - **Developed by:** [More Information Needed]
26
+ - **Funded by [optional]:** [More Information Needed]
27
+ - **Shared by [optional]:** [More Information Needed]
28
+ - **Model type:** [More Information Needed]
29
+ - **Language(s) (NLP):** [More Information Needed]
30
+ - **License:** [More Information Needed]
31
+ - **Finetuned from model [optional]:** [More Information Needed]
32
+
33
+ ### Model Sources [optional]
34
+
35
+ <!-- Provide the basic links for the model. -->
36
+
37
+ - **Repository:** [More Information Needed]
38
+ - **Paper [optional]:** [More Information Needed]
39
+ - **Demo [optional]:** [More Information Needed]
40
+
41
+ ## Uses
42
+
43
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
44
+
45
+ ### Direct Use
46
+
47
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
48
+
49
+ [More Information Needed]
50
+
51
+ ### Downstream Use [optional]
52
+
53
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
54
+
55
+ [More Information Needed]
56
+
57
+ ### Out-of-Scope Use
58
+
59
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
60
+
61
+ [More Information Needed]
62
+
63
+ ## Bias, Risks, and Limitations
64
+
65
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
66
+
67
+ [More Information Needed]
68
+
69
+ ### Recommendations
70
+
71
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
72
+
73
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
74
+
75
+ ## How to Get Started with the Model
76
+
77
+ Use the code below to get started with the model.
78
+
79
+ [More Information Needed]
80
+
81
+ ## Training Details
82
+
83
+ ### Training Data
84
+
85
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
86
+
87
+ [More Information Needed]
88
+
89
+ ### Training Procedure
90
+
91
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
92
+
93
+ #### Preprocessing [optional]
94
+
95
+ [More Information Needed]
96
+
97
+
98
+ #### Training Hyperparameters
99
+
100
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
101
+
102
+ #### Speeds, Sizes, Times [optional]
103
+
104
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
105
+
106
+ [More Information Needed]
107
+
108
+ ## Evaluation
109
+
110
+ <!-- This section describes the evaluation protocols and provides the results. -->
111
+
112
+ ### Testing Data, Factors & Metrics
113
+
114
+ #### Testing Data
115
+
116
+ <!-- This should link to a Dataset Card if possible. -->
117
+
118
+ [More Information Needed]
119
+
120
+ #### Factors
121
+
122
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
123
+
124
+ [More Information Needed]
125
+
126
+ #### Metrics
127
+
128
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
129
+
130
+ [More Information Needed]
131
+
132
+ ### Results
133
+
134
+ [More Information Needed]
135
+
136
+ #### Summary
137
+
138
+
139
+
140
+ ## Model Examination [optional]
141
+
142
+ <!-- Relevant interpretability work for the model goes here -->
143
+
144
+ [More Information Needed]
145
+
146
+ ## Environmental Impact
147
+
148
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
149
+
150
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
151
+
152
+ - **Hardware Type:** [More Information Needed]
153
+ - **Hours used:** [More Information Needed]
154
+ - **Cloud Provider:** [More Information Needed]
155
+ - **Compute Region:** [More Information Needed]
156
+ - **Carbon Emitted:** [More Information Needed]
157
+
158
+ ## Technical Specifications [optional]
159
+
160
+ ### Model Architecture and Objective
161
+
162
+ [More Information Needed]
163
+
164
+ ### Compute Infrastructure
165
+
166
+ [More Information Needed]
167
+
168
+ #### Hardware
169
+
170
+ [More Information Needed]
171
+
172
+ #### Software
173
+
174
+ [More Information Needed]
175
+
176
+ ## Citation [optional]
177
+
178
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
179
+
180
+ **BibTeX:**
181
+
182
+ [More Information Needed]
183
+
184
+ **APA:**
185
+
186
+ [More Information Needed]
187
+
188
+ ## Glossary [optional]
189
+
190
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
191
+
192
+ [More Information Needed]
193
+
194
+ ## More Information [optional]
195
+
196
+ [More Information Needed]
197
+
198
+ ## Model Card Authors [optional]
199
+
200
+ [More Information Needed]
201
+
202
+ ## Model Card Contact
203
+
204
+ [More Information Needed]
205
+ ### Framework versions
206
+
207
+ - PEFT 0.18.1
llama3.1-8b/dp3/adapter/adapter_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "meta-llama/Llama-3.1-8B",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": true,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": [
25
+ "lm_head",
26
+ "embed_tokens"
27
+ ],
28
+ "peft_type": "LORA",
29
+ "peft_version": "0.18.1",
30
+ "qalora_group_size": 16,
31
+ "r": 16,
32
+ "rank_pattern": {},
33
+ "revision": null,
34
+ "target_modules": [
35
+ "q_proj",
36
+ "k_proj",
37
+ "o_proj",
38
+ "v_proj"
39
+ ],
40
+ "target_parameters": null,
41
+ "task_type": "CAUSAL_LM",
42
+ "trainable_token_indices": null,
43
+ "use_dora": false,
44
+ "use_qalora": false,
45
+ "use_rslora": false
46
+ }
llama3.1-8b/dp3/audit_results.json ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "delta": 1e-05,
3
+ "num_canaries": 500,
4
+ "num_members": 250,
5
+ "paper_guess_fraction": 0.2,
6
+ "paper_guess_steps": 20,
7
+ "loss": {
8
+ "auc": 0.505432,
9
+ "empirical_epsilon": {
10
+ "0.05": 0.0,
11
+ "0.01": 0.0
12
+ },
13
+ "empirical_epsilon_details": {
14
+ "0.05": {
15
+ "epsilon": 0.0,
16
+ "num_guesses": 0,
17
+ "correct_guesses": 0,
18
+ "candidate_num_guesses": [
19
+ 5,
20
+ 10,
21
+ 15,
22
+ 20,
23
+ 25,
24
+ 30,
25
+ 35,
26
+ 40,
27
+ 45,
28
+ 50,
29
+ 55,
30
+ 60,
31
+ 65,
32
+ 70,
33
+ 75,
34
+ 80,
35
+ 85,
36
+ 90,
37
+ 95,
38
+ 100
39
+ ],
40
+ "direction": "lower"
41
+ },
42
+ "0.01": {
43
+ "epsilon": 0.0,
44
+ "num_guesses": 0,
45
+ "correct_guesses": 0,
46
+ "candidate_num_guesses": [
47
+ 5,
48
+ 10,
49
+ 15,
50
+ 20,
51
+ 25,
52
+ 30,
53
+ 35,
54
+ 40,
55
+ 45,
56
+ 50,
57
+ 55,
58
+ 60,
59
+ 65,
60
+ 70,
61
+ 75,
62
+ 80,
63
+ 85,
64
+ 90,
65
+ 95,
66
+ 100
67
+ ],
68
+ "direction": "lower"
69
+ }
70
+ }
71
+ },
72
+ "embedding": {
73
+ "auc": 0.50372,
74
+ "empirical_epsilon": {
75
+ "0.05": 0.0,
76
+ "0.01": 0.0
77
+ },
78
+ "empirical_epsilon_details": {
79
+ "0.05": {
80
+ "epsilon": 0.0,
81
+ "num_guesses": 0,
82
+ "correct_guesses": 0,
83
+ "candidate_num_guesses": [
84
+ 5,
85
+ 10,
86
+ 15,
87
+ 20,
88
+ 25,
89
+ 30,
90
+ 35,
91
+ 40,
92
+ 45,
93
+ 50,
94
+ 55,
95
+ 60,
96
+ 65,
97
+ 70,
98
+ 75,
99
+ 80,
100
+ 85,
101
+ 90,
102
+ 95,
103
+ 100
104
+ ],
105
+ "direction": "lower"
106
+ },
107
+ "0.01": {
108
+ "epsilon": 0.0,
109
+ "num_guesses": 0,
110
+ "correct_guesses": 0,
111
+ "candidate_num_guesses": [
112
+ 5,
113
+ 10,
114
+ 15,
115
+ 20,
116
+ 25,
117
+ 30,
118
+ 35,
119
+ 40,
120
+ 45,
121
+ 50,
122
+ 55,
123
+ 60,
124
+ 65,
125
+ 70,
126
+ 75,
127
+ 80,
128
+ 85,
129
+ 90,
130
+ 95,
131
+ 100
132
+ ],
133
+ "direction": "lower"
134
+ }
135
+ }
136
+ }
137
+ }
llama3.1-8b/dp3/canary_meta.json ADDED
The diff for this file is too large to render. See raw diff
 
llama3.1-8b/dp3/metrics.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
llama3.1-8b/dp3/resolved_config.yaml ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ name: meta-llama/Llama-3.1-8B
3
+ tokenizer_name: meta-llama/Llama-3.1-8B
4
+ max_length: 1024
5
+ dtype: bfloat16
6
+ trust_remote_code: true
7
+ use_fast_tokenizer: true
8
+ cache_dir: null
9
+ local_files_only: false
10
+ low_cpu_mem_usage: true
11
+ tie_word_embeddings: true
12
+ gradient_checkpointing: false
13
+ use_chat_template: false
14
+ dataset:
15
+ name: melihcatal/codedp-cpt
16
+ split: train
17
+ mode: cpt
18
+ text_column: text
19
+ validation_ratio: 0.05
20
+ max_samples: -1
21
+ lora:
22
+ enabled: true
23
+ r: 16
24
+ alpha: 32
25
+ dropout: 0.05
26
+ target_modules:
27
+ - q_proj
28
+ - k_proj
29
+ - v_proj
30
+ - o_proj
31
+ modules_to_save:
32
+ - lm_head
33
+ bias: none
34
+ training:
35
+ seed: 42
36
+ epochs: 5
37
+ warmup_steps: null
38
+ warmup_ratio: 0.05
39
+ mixed_precision: false
40
+ mixed_precision_dtype: bfloat16
41
+ batch_size: 4
42
+ eval_batch_size: 8
43
+ eval_every_steps: 10
44
+ eval_every_epochs: 1
45
+ learning_rate: 0.0005
46
+ optimizer: adamw
47
+ lr_scheduler: cosine
48
+ adam_beta1: 0.9
49
+ adam_beta2: 0.999
50
+ adam_epsilon: 1.0e-08
51
+ sgd_momentum: 0.9
52
+ weight_decay: 0.01
53
+ max_grad_norm: 1.0
54
+ log_every: 5
55
+ gradient_accumulation_steps: 8
56
+ num_workers: 4
57
+ output_dir: /scratch/mcatal/runs/cpt/llama3.1-8b/dp3
58
+ min_lr_ratio: 0.15
59
+ distributed:
60
+ strategy: dpddp
61
+ backend: nccl
62
+ devices: null
63
+ dp:
64
+ module_validator: auto
65
+ target_delta: 1.0e-05
66
+ noise_multiplier: null
67
+ max_grad_norm: 1.0
68
+ grad_sample_mode: hooks
69
+ clipping: per_layer
70
+ secure_mode: false
71
+ enabled: true
72
+ target_epsilon: 3.0
73
+ audit:
74
+ enabled: true
75
+ run_every_epoch: true
76
+ epoch_device: cuda
77
+ q_canary: auto
78
+ num_canaries: 500
79
+ prefix_length: 49
80
+ num_digits: 12
81
+ batch_size: 32
82
+ delta: 1.0e-05
83
+ p_values:
84
+ - 0.05
85
+ - 0.01
86
+ paper_guess_fraction: 0.2
87
+ paper_guess_steps: 20
88
+ enable_holdout_empirical_epsilon: false
89
+ holdout_seed: 42
90
+ tie_seed: 42
91
+ tracking:
92
+ enabled: true
93
+ tensorboard: true
94
+ wandb: false
95
+ wandb_project: codedp-finetune-h200-audit
96
+ wandb_run_name: llama3.1-8b-cpt-dp3
97
+ wandb_mode: online
98
+ codecarbon: true
99
+ codecarbon_output_file: codecarbon.csv
100
+ codecarbon_measure_power_secs: 15
101
+ codecarbon_country_iso_code: null
102
+ codecarbon_project_name: codedp-llama3.1-8b-cpt-dp3
llama3.1-8b/dp3/summary.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audit/delta": 1e-05,
3
+ "audit/embedding/auc": 0.50372,
4
+ "audit/embedding/empirical_epsilon/0.01": 0.0,
5
+ "audit/embedding/empirical_epsilon/0.05": 0.0,
6
+ "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 0.0,
7
+ "audit/embedding/empirical_epsilon_details/0.01/epsilon": 0.0,
8
+ "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 0.0,
9
+ "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 0.0,
10
+ "audit/embedding/empirical_epsilon_details/0.05/epsilon": 0.0,
11
+ "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 0.0,
12
+ "audit/loss/auc": 0.505432,
13
+ "audit/loss/empirical_epsilon/0.01": 0.0,
14
+ "audit/loss/empirical_epsilon/0.05": 0.0,
15
+ "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 0.0,
16
+ "audit/loss/empirical_epsilon_details/0.01/epsilon": 0.0,
17
+ "audit/loss/empirical_epsilon_details/0.01/num_guesses": 0.0,
18
+ "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 0.0,
19
+ "audit/loss/empirical_epsilon_details/0.05/epsilon": 0.0,
20
+ "audit/loss/empirical_epsilon_details/0.05/num_guesses": 0.0,
21
+ "audit/num_canaries": 500.0,
22
+ "audit/num_members": 250.0,
23
+ "audit/paper_guess_fraction": 0.2,
24
+ "audit/paper_guess_steps": 20.0,
25
+ "energy/codecarbon/cpu_count": 16.0,
26
+ "energy/codecarbon/cpu_energy": 0.20119429849070633,
27
+ "energy/codecarbon/cpu_power": 80.03162828867059,
28
+ "energy/codecarbon/cpu_utilization_percent": 3.588619282800086,
29
+ "energy/codecarbon/duration": 9403.99333824683,
30
+ "energy/codecarbon/emissions": 0.32646861036645725,
31
+ "energy/codecarbon/emissions_rate": 3.471595508672704e-05,
32
+ "energy/codecarbon/energy_consumed": 9.369704398773274,
33
+ "energy/codecarbon/gpu_count": 8.0,
34
+ "energy/codecarbon/gpu_energy": 9.072981345879256,
35
+ "energy/codecarbon/gpu_power": 3483.890131638086,
36
+ "energy/codecarbon/gpu_utilization_percent": 95.65199183936433,
37
+ "energy/codecarbon/latitude": 47.4843,
38
+ "energy/codecarbon/longitude": 8.212,
39
+ "energy/codecarbon/pue": 1.0,
40
+ "energy/codecarbon/ram_energy": 0.0955287544033269,
41
+ "energy/codecarbon/ram_power": 38.0,
42
+ "energy/codecarbon/ram_total_size": 128.0,
43
+ "energy/codecarbon/ram_used_gb": 509.321770667825,
44
+ "energy/codecarbon/ram_utilization_percent": 25.710800944814256,
45
+ "energy/codecarbon/water_consumed": 0.0,
46
+ "energy/codecarbon/wue": 0.0,
47
+ "eval/duration_sec": 13.896897186990827,
48
+ "eval/loss": 8.814481154466286,
49
+ "perf/audit_duration_sec": 6.752107602078468,
50
+ "perf/epoch_duration_sec": 1849.8364193108864,
51
+ "perf/epoch_samples": 51412.0,
52
+ "perf/epoch_samples_per_sec": 27.792727758680602,
53
+ "perf/epoch_tokens": 37419500.0,
54
+ "perf/epoch_tokens_per_sec": 20228.54540507953,
55
+ "perf/logical_batch_size": 3.0,
56
+ "perf/logical_token_count": 432.0,
57
+ "perf/physical_batches": 2.0,
58
+ "perf/samples_per_sec": 1.6812982665213132,
59
+ "perf/step_duration_sec": 1.7843353911302984,
60
+ "perf/tokens_per_sec": 242.1069503790691,
61
+ "privacy/epsilon": 2.991647548922312,
62
+ "system/cuda_epoch_peak_memory_gb": 62.821791648864746,
63
+ "system/cuda_max_memory_allocated_gb": 62.821791648864746,
64
+ "system/cuda_memory_allocated_gb": 20.329766750335693,
65
+ "train/epoch_canary_loss": 10.630910466309556,
66
+ "train/epoch_loss": 9.228892510061623,
67
+ "train/epoch_real_loss": 8.99651010269704,
68
+ "train/lr": 7.5e-05,
69
+ "train/step_canary_loss": 9.770833333333334,
70
+ "train/step_loss": 9.552105188369751,
71
+ "train/step_real_loss": 8.895918846130371
72
+ }
llama3.1-8b/dp3/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,516 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "bos_token": "<|begin_of_text|>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|end_of_text|>",
6
+ "extra_special_tokens": {
7
+ "extra_special_token_0": "865331112869",
8
+ "extra_special_token_1": "569765693871",
9
+ "extra_special_token_2": "485177821815",
10
+ "extra_special_token_3": "135441121756",
11
+ "extra_special_token_4": "367459894796",
12
+ "extra_special_token_5": "877482678543",
13
+ "extra_special_token_6": "457919547633",
14
+ "extra_special_token_7": "765474393376",
15
+ "extra_special_token_8": "114848338811",
16
+ "extra_special_token_9": "746285987371",
17
+ "extra_special_token_10": "649291669397",
18
+ "extra_special_token_11": "927914615679",
19
+ "extra_special_token_12": "445925149649",
20
+ "extra_special_token_13": "691587454538",
21
+ "extra_special_token_14": "143777992227",
22
+ "extra_special_token_15": "997981281989",
23
+ "extra_special_token_16": "425949483533",
24
+ "extra_special_token_17": "982993456429",
25
+ "extra_special_token_18": "718726519731",
26
+ "extra_special_token_19": "172599315861",
27
+ "extra_special_token_20": "643489267333",
28
+ "extra_special_token_21": "282322838685",
29
+ "extra_special_token_22": "781653545886",
30
+ "extra_special_token_23": "796415361892",
31
+ "extra_special_token_24": "841991688488",
32
+ "extra_special_token_25": "211411365397",
33
+ "extra_special_token_26": "698218415444",
34
+ "extra_special_token_27": "355977139358",
35
+ "extra_special_token_28": "682564697312",
36
+ "extra_special_token_29": "383837596997",
37
+ "extra_special_token_30": "689362171782",
38
+ "extra_special_token_31": "749966767285",
39
+ "extra_special_token_32": "753159165157",
40
+ "extra_special_token_33": "795693824762",
41
+ "extra_special_token_34": "669689115557",
42
+ "extra_special_token_35": "327491773134",
43
+ "extra_special_token_36": "983569279932",
44
+ "extra_special_token_37": "612128769512",
45
+ "extra_special_token_38": "374327157578",
46
+ "extra_special_token_39": "311632789559",
47
+ "extra_special_token_40": "523918658846",
48
+ "extra_special_token_41": "765981581453",
49
+ "extra_special_token_42": "794825141891",
50
+ "extra_special_token_43": "873898736873",
51
+ "extra_special_token_44": "447445629421",
52
+ "extra_special_token_45": "473822473819",
53
+ "extra_special_token_46": "181439694557",
54
+ "extra_special_token_47": "592538279337",
55
+ "extra_special_token_48": "668134915514",
56
+ "extra_special_token_49": "643692393748",
57
+ "extra_special_token_50": "696651276628",
58
+ "extra_special_token_51": "853859348234",
59
+ "extra_special_token_52": "778466723723",
60
+ "extra_special_token_53": "929826356991",
61
+ "extra_special_token_54": "272362973463",
62
+ "extra_special_token_55": "694235616268",
63
+ "extra_special_token_56": "281673864127",
64
+ "extra_special_token_57": "479676316326",
65
+ "extra_special_token_58": "646979124677",
66
+ "extra_special_token_59": "922327493433",
67
+ "extra_special_token_60": "883685933161",
68
+ "extra_special_token_61": "264259917554",
69
+ "extra_special_token_62": "836746273134",
70
+ "extra_special_token_63": "658481324922",
71
+ "extra_special_token_64": "481884157827",
72
+ "extra_special_token_65": "587787496812",
73
+ "extra_special_token_66": "579184949249",
74
+ "extra_special_token_67": "912193598348",
75
+ "extra_special_token_68": "529679678956",
76
+ "extra_special_token_69": "795838284624",
77
+ "extra_special_token_70": "159337222655",
78
+ "extra_special_token_71": "173781362446",
79
+ "extra_special_token_72": "773687856563",
80
+ "extra_special_token_73": "535787224917",
81
+ "extra_special_token_74": "351885857332",
82
+ "extra_special_token_75": "578827344666",
83
+ "extra_special_token_76": "198462689911",
84
+ "extra_special_token_77": "722618266242",
85
+ "extra_special_token_78": "952872416512",
86
+ "extra_special_token_79": "517778845323",
87
+ "extra_special_token_80": "749665846687",
88
+ "extra_special_token_81": "661436365453",
89
+ "extra_special_token_82": "259666844669",
90
+ "extra_special_token_83": "242851284913",
91
+ "extra_special_token_84": "514532995959",
92
+ "extra_special_token_85": "161588262349",
93
+ "extra_special_token_86": "742765629356",
94
+ "extra_special_token_87": "225164373623",
95
+ "extra_special_token_88": "676539973863",
96
+ "extra_special_token_89": "826214551218",
97
+ "extra_special_token_90": "182345464792",
98
+ "extra_special_token_91": "232776999554",
99
+ "extra_special_token_92": "337326533813",
100
+ "extra_special_token_93": "676676697292",
101
+ "extra_special_token_94": "929185622831",
102
+ "extra_special_token_95": "545512344383",
103
+ "extra_special_token_96": "499444466686",
104
+ "extra_special_token_97": "314697386682",
105
+ "extra_special_token_98": "517379856925",
106
+ "extra_special_token_99": "379557332953",
107
+ "extra_special_token_100": "614797267726",
108
+ "extra_special_token_101": "429781429464",
109
+ "extra_special_token_102": "922466849763",
110
+ "extra_special_token_103": "721737645236",
111
+ "extra_special_token_104": "479227349997",
112
+ "extra_special_token_105": "136931728327",
113
+ "extra_special_token_106": "259533577263",
114
+ "extra_special_token_107": "488538864842",
115
+ "extra_special_token_108": "937495658852",
116
+ "extra_special_token_109": "489991411364",
117
+ "extra_special_token_110": "499148455254",
118
+ "extra_special_token_111": "441373944925",
119
+ "extra_special_token_112": "899151413682",
120
+ "extra_special_token_113": "467893531755",
121
+ "extra_special_token_114": "527117488925",
122
+ "extra_special_token_115": "928335588653",
123
+ "extra_special_token_116": "374439448821",
124
+ "extra_special_token_117": "879425227932",
125
+ "extra_special_token_118": "867678158885",
126
+ "extra_special_token_119": "399749397872",
127
+ "extra_special_token_120": "129693547287",
128
+ "extra_special_token_121": "689285841825",
129
+ "extra_special_token_122": "771619544974",
130
+ "extra_special_token_123": "724883568652",
131
+ "extra_special_token_124": "516968424863",
132
+ "extra_special_token_125": "733737988257",
133
+ "extra_special_token_126": "852347289392",
134
+ "extra_special_token_127": "296953381169",
135
+ "extra_special_token_128": "377273562477",
136
+ "extra_special_token_129": "262296912232",
137
+ "extra_special_token_130": "547149832394",
138
+ "extra_special_token_131": "298464134954",
139
+ "extra_special_token_132": "216667245274",
140
+ "extra_special_token_133": "843998562287",
141
+ "extra_special_token_134": "572154333646",
142
+ "extra_special_token_135": "124589118494",
143
+ "extra_special_token_136": "841824384614",
144
+ "extra_special_token_137": "232896526252",
145
+ "extra_special_token_138": "295448593321",
146
+ "extra_special_token_139": "123741461297",
147
+ "extra_special_token_140": "653573457168",
148
+ "extra_special_token_141": "196735786156",
149
+ "extra_special_token_142": "377338713663",
150
+ "extra_special_token_143": "964342468552",
151
+ "extra_special_token_144": "586855179568",
152
+ "extra_special_token_145": "484773717614",
153
+ "extra_special_token_146": "894885246797",
154
+ "extra_special_token_147": "677896358599",
155
+ "extra_special_token_148": "848845611563",
156
+ "extra_special_token_149": "851852651677",
157
+ "extra_special_token_150": "398549545767",
158
+ "extra_special_token_151": "454244839926",
159
+ "extra_special_token_152": "799364566435",
160
+ "extra_special_token_153": "967114116556",
161
+ "extra_special_token_154": "817378986438",
162
+ "extra_special_token_155": "233795848681",
163
+ "extra_special_token_156": "824387273757",
164
+ "extra_special_token_157": "916198946615",
165
+ "extra_special_token_158": "563117729724",
166
+ "extra_special_token_159": "951794811935",
167
+ "extra_special_token_160": "374598961236",
168
+ "extra_special_token_161": "922867396683",
169
+ "extra_special_token_162": "765737843639",
170
+ "extra_special_token_163": "175469284871",
171
+ "extra_special_token_164": "231853711778",
172
+ "extra_special_token_165": "662426712668",
173
+ "extra_special_token_166": "711412347158",
174
+ "extra_special_token_167": "753466987363",
175
+ "extra_special_token_168": "513361312532",
176
+ "extra_special_token_169": "712992815957",
177
+ "extra_special_token_170": "971621888444",
178
+ "extra_special_token_171": "829235161526",
179
+ "extra_special_token_172": "585544633356",
180
+ "extra_special_token_173": "582471228164",
181
+ "extra_special_token_174": "678666359123",
182
+ "extra_special_token_175": "557533689478",
183
+ "extra_special_token_176": "632962475133",
184
+ "extra_special_token_177": "484489193824",
185
+ "extra_special_token_178": "489562189822",
186
+ "extra_special_token_179": "589547936288",
187
+ "extra_special_token_180": "363214487524",
188
+ "extra_special_token_181": "244885399387",
189
+ "extra_special_token_182": "431751228368",
190
+ "extra_special_token_183": "433581868192",
191
+ "extra_special_token_184": "486391569221",
192
+ "extra_special_token_185": "185438575221",
193
+ "extra_special_token_186": "126574388585",
194
+ "extra_special_token_187": "741757479784",
195
+ "extra_special_token_188": "529854679937",
196
+ "extra_special_token_189": "996116119839",
197
+ "extra_special_token_190": "616248973917",
198
+ "extra_special_token_191": "763531783491",
199
+ "extra_special_token_192": "955456118295",
200
+ "extra_special_token_193": "364196983365",
201
+ "extra_special_token_194": "195792996468",
202
+ "extra_special_token_195": "151859598873",
203
+ "extra_special_token_196": "399223169721",
204
+ "extra_special_token_197": "938488813964",
205
+ "extra_special_token_198": "961981959227",
206
+ "extra_special_token_199": "183368827562",
207
+ "extra_special_token_200": "533417736566",
208
+ "extra_special_token_201": "786391632558",
209
+ "extra_special_token_202": "665661658354",
210
+ "extra_special_token_203": "693281533643",
211
+ "extra_special_token_204": "475794684356",
212
+ "extra_special_token_205": "652154162978",
213
+ "extra_special_token_206": "753233719644",
214
+ "extra_special_token_207": "668514843129",
215
+ "extra_special_token_208": "819162623892",
216
+ "extra_special_token_209": "941169431859",
217
+ "extra_special_token_210": "877385381798",
218
+ "extra_special_token_211": "752644929761",
219
+ "extra_special_token_212": "881136466196",
220
+ "extra_special_token_213": "275597777299",
221
+ "extra_special_token_214": "731681792655",
222
+ "extra_special_token_215": "961133895172",
223
+ "extra_special_token_216": "864718285734",
224
+ "extra_special_token_217": "963852916563",
225
+ "extra_special_token_218": "319584985416",
226
+ "extra_special_token_219": "563365646341",
227
+ "extra_special_token_220": "811371928234",
228
+ "extra_special_token_221": "837131396371",
229
+ "extra_special_token_222": "267514771964",
230
+ "extra_special_token_223": "944513428457",
231
+ "extra_special_token_224": "117298239631",
232
+ "extra_special_token_225": "158142752582",
233
+ "extra_special_token_226": "252867443568",
234
+ "extra_special_token_227": "839269684865",
235
+ "extra_special_token_228": "612788593128",
236
+ "extra_special_token_229": "145669731981",
237
+ "extra_special_token_230": "121557291859",
238
+ "extra_special_token_231": "245416776926",
239
+ "extra_special_token_232": "799417897197",
240
+ "extra_special_token_233": "997958836435",
241
+ "extra_special_token_234": "892336777248",
242
+ "extra_special_token_235": "158929292238",
243
+ "extra_special_token_236": "581976444672",
244
+ "extra_special_token_237": "897784492783",
245
+ "extra_special_token_238": "492373714791",
246
+ "extra_special_token_239": "512659818733",
247
+ "extra_special_token_240": "881112998642",
248
+ "extra_special_token_241": "619454958782",
249
+ "extra_special_token_242": "431149748713",
250
+ "extra_special_token_243": "624221476921",
251
+ "extra_special_token_244": "125866399464",
252
+ "extra_special_token_245": "339882449689",
253
+ "extra_special_token_246": "186198784585",
254
+ "extra_special_token_247": "943193294691",
255
+ "extra_special_token_248": "955668961269",
256
+ "extra_special_token_249": "232787996724",
257
+ "extra_special_token_250": "215671314196",
258
+ "extra_special_token_251": "286173241916",
259
+ "extra_special_token_252": "745977673725",
260
+ "extra_special_token_253": "556976448182",
261
+ "extra_special_token_254": "599961512792",
262
+ "extra_special_token_255": "766294538337",
263
+ "extra_special_token_256": "934912591213",
264
+ "extra_special_token_257": "295118729589",
265
+ "extra_special_token_258": "529455466433",
266
+ "extra_special_token_259": "196119929397",
267
+ "extra_special_token_260": "379571934299",
268
+ "extra_special_token_261": "251789649997",
269
+ "extra_special_token_262": "564544131355",
270
+ "extra_special_token_263": "244371196654",
271
+ "extra_special_token_264": "384598329253",
272
+ "extra_special_token_265": "887753195844",
273
+ "extra_special_token_266": "364947325679",
274
+ "extra_special_token_267": "655517954651",
275
+ "extra_special_token_268": "673948786567",
276
+ "extra_special_token_269": "857231548835",
277
+ "extra_special_token_270": "816115936673",
278
+ "extra_special_token_271": "644234165531",
279
+ "extra_special_token_272": "182782912224",
280
+ "extra_special_token_273": "234316622259",
281
+ "extra_special_token_274": "421369185549",
282
+ "extra_special_token_275": "434632855397",
283
+ "extra_special_token_276": "921889371893",
284
+ "extra_special_token_277": "415956914763",
285
+ "extra_special_token_278": "598916996413",
286
+ "extra_special_token_279": "773671349113",
287
+ "extra_special_token_280": "952465217972",
288
+ "extra_special_token_281": "117657531962",
289
+ "extra_special_token_282": "729825168745",
290
+ "extra_special_token_283": "691315125346",
291
+ "extra_special_token_284": "768461952319",
292
+ "extra_special_token_285": "664847713559",
293
+ "extra_special_token_286": "953267689786",
294
+ "extra_special_token_287": "886464195129",
295
+ "extra_special_token_288": "824488329416",
296
+ "extra_special_token_289": "837873762491",
297
+ "extra_special_token_290": "532833541879",
298
+ "extra_special_token_291": "669183782449",
299
+ "extra_special_token_292": "941976537588",
300
+ "extra_special_token_293": "739394546916",
301
+ "extra_special_token_294": "267954879268",
302
+ "extra_special_token_295": "637551427887",
303
+ "extra_special_token_296": "217756494954",
304
+ "extra_special_token_297": "524444658383",
305
+ "extra_special_token_298": "117783274348",
306
+ "extra_special_token_299": "138218735276",
307
+ "extra_special_token_300": "814611949491",
308
+ "extra_special_token_301": "711641973413",
309
+ "extra_special_token_302": "499156317423",
310
+ "extra_special_token_303": "515856611931",
311
+ "extra_special_token_304": "454164859837",
312
+ "extra_special_token_305": "345271433112",
313
+ "extra_special_token_306": "462294118988",
314
+ "extra_special_token_307": "511785788222",
315
+ "extra_special_token_308": "497294727353",
316
+ "extra_special_token_309": "866519986723",
317
+ "extra_special_token_310": "334513529294",
318
+ "extra_special_token_311": "549946382131",
319
+ "extra_special_token_312": "284445431422",
320
+ "extra_special_token_313": "396521188476",
321
+ "extra_special_token_314": "421435255895",
322
+ "extra_special_token_315": "133373659361",
323
+ "extra_special_token_316": "322683334381",
324
+ "extra_special_token_317": "228358422847",
325
+ "extra_special_token_318": "291762694874",
326
+ "extra_special_token_319": "143182978129",
327
+ "extra_special_token_320": "511923256573",
328
+ "extra_special_token_321": "327158398268",
329
+ "extra_special_token_322": "879764613759",
330
+ "extra_special_token_323": "564395222747",
331
+ "extra_special_token_324": "451161679736",
332
+ "extra_special_token_325": "538631466654",
333
+ "extra_special_token_326": "221762325616",
334
+ "extra_special_token_327": "218391991184",
335
+ "extra_special_token_328": "322589379462",
336
+ "extra_special_token_329": "876537814263",
337
+ "extra_special_token_330": "152676556624",
338
+ "extra_special_token_331": "332522971941",
339
+ "extra_special_token_332": "884354318946",
340
+ "extra_special_token_333": "513349618943",
341
+ "extra_special_token_334": "116639746413",
342
+ "extra_special_token_335": "635185846287",
343
+ "extra_special_token_336": "993832498489",
344
+ "extra_special_token_337": "813981174797",
345
+ "extra_special_token_338": "438745114173",
346
+ "extra_special_token_339": "983493951323",
347
+ "extra_special_token_340": "724492262421",
348
+ "extra_special_token_341": "622553389126",
349
+ "extra_special_token_342": "889965243135",
350
+ "extra_special_token_343": "364492359246",
351
+ "extra_special_token_344": "154962668224",
352
+ "extra_special_token_345": "179564995814",
353
+ "extra_special_token_346": "418412875665",
354
+ "extra_special_token_347": "718951851413",
355
+ "extra_special_token_348": "699446724178",
356
+ "extra_special_token_349": "624266421831",
357
+ "extra_special_token_350": "815458725125",
358
+ "extra_special_token_351": "455423278865",
359
+ "extra_special_token_352": "393741199486",
360
+ "extra_special_token_353": "328552864359",
361
+ "extra_special_token_354": "211662639865",
362
+ "extra_special_token_355": "218784516525",
363
+ "extra_special_token_356": "762486672996",
364
+ "extra_special_token_357": "142799718159",
365
+ "extra_special_token_358": "858146415154",
366
+ "extra_special_token_359": "767858144912",
367
+ "extra_special_token_360": "571317457151",
368
+ "extra_special_token_361": "635127952696",
369
+ "extra_special_token_362": "116427191984",
370
+ "extra_special_token_363": "268921994538",
371
+ "extra_special_token_364": "523937669294",
372
+ "extra_special_token_365": "165429152138",
373
+ "extra_special_token_366": "739246183345",
374
+ "extra_special_token_367": "591464355756",
375
+ "extra_special_token_368": "212985874612",
376
+ "extra_special_token_369": "191887635211",
377
+ "extra_special_token_370": "967214577653",
378
+ "extra_special_token_371": "119342152414",
379
+ "extra_special_token_372": "946444632795",
380
+ "extra_special_token_373": "618423867817",
381
+ "extra_special_token_374": "228565148417",
382
+ "extra_special_token_375": "729116422489",
383
+ "extra_special_token_376": "527874729936",
384
+ "extra_special_token_377": "739784153482",
385
+ "extra_special_token_378": "387763951128",
386
+ "extra_special_token_379": "331369926711",
387
+ "extra_special_token_380": "562716493614",
388
+ "extra_special_token_381": "739667844957",
389
+ "extra_special_token_382": "562389434565",
390
+ "extra_special_token_383": "256497188281",
391
+ "extra_special_token_384": "859927364588",
392
+ "extra_special_token_385": "417668946583",
393
+ "extra_special_token_386": "357621613582",
394
+ "extra_special_token_387": "438435178228",
395
+ "extra_special_token_388": "485692541169",
396
+ "extra_special_token_389": "825815739116",
397
+ "extra_special_token_390": "342221452223",
398
+ "extra_special_token_391": "697747991249",
399
+ "extra_special_token_392": "716763689965",
400
+ "extra_special_token_393": "141499982867",
401
+ "extra_special_token_394": "818479319499",
402
+ "extra_special_token_395": "336813343298",
403
+ "extra_special_token_396": "594688742928",
404
+ "extra_special_token_397": "472129283475",
405
+ "extra_special_token_398": "514354144759",
406
+ "extra_special_token_399": "349249721685",
407
+ "extra_special_token_400": "546276298359",
408
+ "extra_special_token_401": "353755529131",
409
+ "extra_special_token_402": "315534574435",
410
+ "extra_special_token_403": "523723475786",
411
+ "extra_special_token_404": "215826764872",
412
+ "extra_special_token_405": "367968398551",
413
+ "extra_special_token_406": "569853653352",
414
+ "extra_special_token_407": "389715484387",
415
+ "extra_special_token_408": "293847485454",
416
+ "extra_special_token_409": "714738141818",
417
+ "extra_special_token_410": "178478368922",
418
+ "extra_special_token_411": "581493616981",
419
+ "extra_special_token_412": "589439538674",
420
+ "extra_special_token_413": "846657726193",
421
+ "extra_special_token_414": "722339992679",
422
+ "extra_special_token_415": "138154781148",
423
+ "extra_special_token_416": "757785319772",
424
+ "extra_special_token_417": "492516914298",
425
+ "extra_special_token_418": "919181521716",
426
+ "extra_special_token_419": "985781138935",
427
+ "extra_special_token_420": "476969195485",
428
+ "extra_special_token_421": "313145133463",
429
+ "extra_special_token_422": "758963111966",
430
+ "extra_special_token_423": "147541537162",
431
+ "extra_special_token_424": "557163366873",
432
+ "extra_special_token_425": "144373897488",
433
+ "extra_special_token_426": "522515164754",
434
+ "extra_special_token_427": "724964923582",
435
+ "extra_special_token_428": "284776712475",
436
+ "extra_special_token_429": "375429755114",
437
+ "extra_special_token_430": "181233596124",
438
+ "extra_special_token_431": "948585673431",
439
+ "extra_special_token_432": "243165586174",
440
+ "extra_special_token_433": "396847976144",
441
+ "extra_special_token_434": "997724962668",
442
+ "extra_special_token_435": "558837194455",
443
+ "extra_special_token_436": "163165456396",
444
+ "extra_special_token_437": "378749551722",
445
+ "extra_special_token_438": "161238482259",
446
+ "extra_special_token_439": "754978243758",
447
+ "extra_special_token_440": "195388849133",
448
+ "extra_special_token_441": "229775525672",
449
+ "extra_special_token_442": "262437452884",
450
+ "extra_special_token_443": "441377892146",
451
+ "extra_special_token_444": "451885565366",
452
+ "extra_special_token_445": "981277526855",
453
+ "extra_special_token_446": "762495822823",
454
+ "extra_special_token_447": "368763327262",
455
+ "extra_special_token_448": "757422791351",
456
+ "extra_special_token_449": "636324136426",
457
+ "extra_special_token_450": "214193645583",
458
+ "extra_special_token_451": "412843856172",
459
+ "extra_special_token_452": "179386156569",
460
+ "extra_special_token_453": "756916173536",
461
+ "extra_special_token_454": "892697125149",
462
+ "extra_special_token_455": "625334487352",
463
+ "extra_special_token_456": "941861857715",
464
+ "extra_special_token_457": "887417525236",
465
+ "extra_special_token_458": "649516938598",
466
+ "extra_special_token_459": "717628619782",
467
+ "extra_special_token_460": "438124184139",
468
+ "extra_special_token_461": "547563892268",
469
+ "extra_special_token_462": "856317483891",
470
+ "extra_special_token_463": "313313831273",
471
+ "extra_special_token_464": "371496153876",
472
+ "extra_special_token_465": "587541149322",
473
+ "extra_special_token_466": "265847332563",
474
+ "extra_special_token_467": "449549215429",
475
+ "extra_special_token_468": "163497196769",
476
+ "extra_special_token_469": "861342291298",
477
+ "extra_special_token_470": "268433315926",
478
+ "extra_special_token_471": "774679513717",
479
+ "extra_special_token_472": "851254219729",
480
+ "extra_special_token_473": "583527834464",
481
+ "extra_special_token_474": "488496781997",
482
+ "extra_special_token_475": "556814553861",
483
+ "extra_special_token_476": "482829231639",
484
+ "extra_special_token_477": "618878266619",
485
+ "extra_special_token_478": "147444452794",
486
+ "extra_special_token_479": "949235426629",
487
+ "extra_special_token_480": "357299947518",
488
+ "extra_special_token_481": "175528632226",
489
+ "extra_special_token_482": "645527857972",
490
+ "extra_special_token_483": "186872457894",
491
+ "extra_special_token_484": "552738847828",
492
+ "extra_special_token_485": "626748382482",
493
+ "extra_special_token_486": "921894985642",
494
+ "extra_special_token_487": "943878645871",
495
+ "extra_special_token_488": "859289776479",
496
+ "extra_special_token_489": "614583493135",
497
+ "extra_special_token_490": "933775286797",
498
+ "extra_special_token_491": "332234613346",
499
+ "extra_special_token_492": "325196781219",
500
+ "extra_special_token_493": "142526557681",
501
+ "extra_special_token_494": "356722692178",
502
+ "extra_special_token_495": "449318681694",
503
+ "extra_special_token_496": "687284547244",
504
+ "extra_special_token_497": "947262995132",
505
+ "extra_special_token_498": "893974619684",
506
+ "extra_special_token_499": "797238311233"
507
+ },
508
+ "is_local": false,
509
+ "model_input_names": [
510
+ "input_ids",
511
+ "attention_mask"
512
+ ],
513
+ "model_max_length": 131072,
514
+ "pad_token": "<|end_of_text|>",
515
+ "tokenizer_class": "PreTrainedTokenizerFast"
516
+ }
llama3.1-8b/dp3/train.log ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-03-29 05:41:19,133 [INFO] new_opacus_codex.train_steps: epoch=1 step=5 loss=12.7584
2
+ 2026-03-29 05:42:01,773 [INFO] new_opacus_codex.train_steps: epoch=1 step=10 loss=12.7560
3
+ 2026-03-29 05:42:15,584 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=10 eval_loss=13.0737 duration_sec=13.80
4
+ 2026-03-29 05:42:58,750 [INFO] new_opacus_codex.train_steps: epoch=1 step=15 loss=12.7511
5
+ 2026-03-29 05:43:41,767 [INFO] new_opacus_codex.train_steps: epoch=1 step=20 loss=12.7096
6
+ 2026-03-29 05:43:55,603 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=20 eval_loss=13.0607 duration_sec=13.82
7
+ 2026-03-29 05:44:38,212 [INFO] new_opacus_codex.train_steps: epoch=1 step=25 loss=12.7708
8
+ 2026-03-29 05:45:20,339 [INFO] new_opacus_codex.train_steps: epoch=1 step=30 loss=12.7645
9
+ 2026-03-29 05:45:34,155 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=30 eval_loss=13.0389 duration_sec=13.80
10
+ 2026-03-29 05:46:16,963 [INFO] new_opacus_codex.train_steps: epoch=1 step=35 loss=12.6553
11
+ 2026-03-29 05:46:58,869 [INFO] new_opacus_codex.train_steps: epoch=1 step=40 loss=12.7876
12
+ 2026-03-29 05:47:12,694 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=40 eval_loss=13.0176 duration_sec=13.81
13
+ 2026-03-29 05:47:55,009 [INFO] new_opacus_codex.train_steps: epoch=1 step=45 loss=12.8207
14
+ 2026-03-29 05:48:36,832 [INFO] new_opacus_codex.train_steps: epoch=1 step=50 loss=12.7124
15
+ 2026-03-29 05:48:50,695 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=50 eval_loss=12.9997 duration_sec=13.85
16
+ 2026-03-29 05:49:32,064 [INFO] new_opacus_codex.train_steps: epoch=1 step=55 loss=12.7316
17
+ 2026-03-29 05:50:14,354 [INFO] new_opacus_codex.train_steps: epoch=1 step=60 loss=12.7261
18
+ 2026-03-29 05:50:28,172 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=60 eval_loss=12.9745 duration_sec=13.81
19
+ 2026-03-29 05:51:10,608 [INFO] new_opacus_codex.train_steps: epoch=1 step=65 loss=12.7628
20
+ 2026-03-29 05:51:52,890 [INFO] new_opacus_codex.train_steps: epoch=1 step=70 loss=12.8104
21
+ 2026-03-29 05:52:06,756 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=70 eval_loss=12.9424 duration_sec=13.85
22
+ 2026-03-29 05:52:49,527 [INFO] new_opacus_codex.train_steps: epoch=1 step=75 loss=12.7921
23
+ 2026-03-29 05:53:33,068 [INFO] new_opacus_codex.train_steps: epoch=1 step=80 loss=12.6775
24
+ 2026-03-29 05:53:46,886 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=80 eval_loss=12.9090 duration_sec=13.81
25
+ 2026-03-29 05:54:29,483 [INFO] new_opacus_codex.train_steps: epoch=1 step=85 loss=12.5991
26
+ 2026-03-29 05:55:12,413 [INFO] new_opacus_codex.train_steps: epoch=1 step=90 loss=12.5933
27
+ 2026-03-29 05:55:26,203 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=90 eval_loss=12.8782 duration_sec=13.78
28
+ 2026-03-29 05:56:08,155 [INFO] new_opacus_codex.train_steps: epoch=1 step=95 loss=12.6907
29
+ 2026-03-29 05:56:50,079 [INFO] new_opacus_codex.train_steps: epoch=1 step=100 loss=12.7816
30
+ 2026-03-29 05:57:03,905 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=100 eval_loss=12.8351 duration_sec=13.81
31
+ 2026-03-29 05:57:47,556 [INFO] new_opacus_codex.train_steps: epoch=1 step=105 loss=12.6913
32
+ 2026-03-29 05:58:29,181 [INFO] new_opacus_codex.train_steps: epoch=1 step=110 loss=12.7056
33
+ 2026-03-29 05:58:43,041 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=110 eval_loss=12.7949 duration_sec=13.85
34
+ 2026-03-29 05:59:25,787 [INFO] new_opacus_codex.train_steps: epoch=1 step=115 loss=12.6525
35
+ 2026-03-29 06:00:07,915 [INFO] new_opacus_codex.train_steps: epoch=1 step=120 loss=12.5624
36
+ 2026-03-29 06:00:21,778 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=120 eval_loss=12.7514 duration_sec=13.85
37
+ 2026-03-29 06:01:03,495 [INFO] new_opacus_codex.train_steps: epoch=1 step=125 loss=12.6408
38
+ 2026-03-29 06:01:45,585 [INFO] new_opacus_codex.train_steps: epoch=1 step=130 loss=12.5745
39
+ 2026-03-29 06:01:59,399 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=130 eval_loss=12.7088 duration_sec=13.80
40
+ 2026-03-29 06:02:41,929 [INFO] new_opacus_codex.train_steps: epoch=1 step=135 loss=12.4956
41
+ 2026-03-29 06:03:23,197 [INFO] new_opacus_codex.train_steps: epoch=1 step=140 loss=12.5767
42
+ 2026-03-29 06:03:37,038 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=140 eval_loss=12.6557 duration_sec=13.83
43
+ 2026-03-29 06:04:19,192 [INFO] new_opacus_codex.train_steps: epoch=1 step=145 loss=12.5427
44
+ 2026-03-29 06:05:02,251 [INFO] new_opacus_codex.train_steps: epoch=1 step=150 loss=12.4583
45
+ 2026-03-29 06:05:16,153 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=150 eval_loss=12.5964 duration_sec=13.89
46
+ 2026-03-29 06:05:57,973 [INFO] new_opacus_codex.train_steps: epoch=1 step=155 loss=12.3189
47
+ 2026-03-29 06:06:38,677 [INFO] new_opacus_codex.train_steps: epoch=1 step=160 loss=12.3426
48
+ 2026-03-29 06:06:52,522 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=160 eval_loss=12.5357 duration_sec=13.83
49
+ 2026-03-29 06:07:35,333 [INFO] new_opacus_codex.train_steps: epoch=1 step=165 loss=12.4614
50
+ 2026-03-29 06:08:18,390 [INFO] new_opacus_codex.train_steps: epoch=1 step=170 loss=12.3697
51
+ 2026-03-29 06:08:32,231 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=170 eval_loss=12.4688 duration_sec=13.83
52
+ 2026-03-29 06:09:14,008 [INFO] new_opacus_codex.train_steps: epoch=1 step=175 loss=12.2808
53
+ 2026-03-29 06:09:57,461 [INFO] new_opacus_codex.train_steps: epoch=1 step=180 loss=12.3229
54
+ 2026-03-29 06:10:11,284 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=180 eval_loss=12.3996 duration_sec=13.81
55
+ 2026-03-29 06:11:15,406 [INFO] new_opacus_codex.train_steps: epoch=2 step=185 loss=12.5593
56
+ 2026-03-29 06:11:58,971 [INFO] new_opacus_codex.train_steps: epoch=2 step=190 loss=12.2930
57
+ 2026-03-29 06:12:12,779 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=190 eval_loss=12.3236 duration_sec=13.79
58
+ 2026-03-29 06:12:55,427 [INFO] new_opacus_codex.train_steps: epoch=2 step=195 loss=12.2362
59
+ 2026-03-29 06:13:38,049 [INFO] new_opacus_codex.train_steps: epoch=2 step=200 loss=12.1393
60
+ 2026-03-29 06:13:51,868 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=200 eval_loss=12.2470 duration_sec=13.81
61
+ 2026-03-29 06:14:33,575 [INFO] new_opacus_codex.train_steps: epoch=2 step=205 loss=12.0464
62
+ 2026-03-29 06:15:16,591 [INFO] new_opacus_codex.train_steps: epoch=2 step=210 loss=12.0315
63
+ 2026-03-29 06:15:30,445 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=210 eval_loss=12.1576 duration_sec=13.84
64
+ 2026-03-29 06:16:13,323 [INFO] new_opacus_codex.train_steps: epoch=2 step=215 loss=12.0206
65
+ 2026-03-29 06:16:57,159 [INFO] new_opacus_codex.train_steps: epoch=2 step=220 loss=12.0425
66
+ 2026-03-29 06:17:11,047 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=220 eval_loss=12.0679 duration_sec=13.88
67
+ 2026-03-29 06:17:54,072 [INFO] new_opacus_codex.train_steps: epoch=2 step=225 loss=11.9556
68
+ 2026-03-29 06:18:36,200 [INFO] new_opacus_codex.train_steps: epoch=2 step=230 loss=11.9417
69
+ 2026-03-29 06:18:50,081 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=230 eval_loss=11.9778 duration_sec=13.87
70
+ 2026-03-29 06:19:32,227 [INFO] new_opacus_codex.train_steps: epoch=2 step=235 loss=12.0425
71
+ 2026-03-29 06:20:15,147 [INFO] new_opacus_codex.train_steps: epoch=2 step=240 loss=12.0188
72
+ 2026-03-29 06:20:29,114 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=240 eval_loss=11.8833 duration_sec=13.95
73
+ 2026-03-29 06:21:11,825 [INFO] new_opacus_codex.train_steps: epoch=2 step=245 loss=11.8227
74
+ 2026-03-29 06:21:54,906 [INFO] new_opacus_codex.train_steps: epoch=2 step=250 loss=11.7866
75
+ 2026-03-29 06:22:08,780 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=250 eval_loss=11.7964 duration_sec=13.86
76
+ 2026-03-29 06:22:52,070 [INFO] new_opacus_codex.train_steps: epoch=2 step=255 loss=11.7588
77
+ 2026-03-29 06:23:33,620 [INFO] new_opacus_codex.train_steps: epoch=2 step=260 loss=11.6342
78
+ 2026-03-29 06:23:47,552 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=260 eval_loss=11.7074 duration_sec=13.92
79
+ 2026-03-29 06:24:29,937 [INFO] new_opacus_codex.train_steps: epoch=2 step=265 loss=11.6222
80
+ 2026-03-29 06:25:11,618 [INFO] new_opacus_codex.train_steps: epoch=2 step=270 loss=11.6388
81
+ 2026-03-29 06:25:25,558 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=270 eval_loss=11.6100 duration_sec=13.93
82
+ 2026-03-29 06:26:08,077 [INFO] new_opacus_codex.train_steps: epoch=2 step=275 loss=11.6365
83
+ 2026-03-29 06:26:52,421 [INFO] new_opacus_codex.train_steps: epoch=2 step=280 loss=11.5211
84
+ 2026-03-29 06:27:06,371 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=280 eval_loss=11.5061 duration_sec=13.93
85
+ 2026-03-29 06:27:48,679 [INFO] new_opacus_codex.train_steps: epoch=2 step=285 loss=11.4539
86
+ 2026-03-29 06:28:31,212 [INFO] new_opacus_codex.train_steps: epoch=2 step=290 loss=11.4306
87
+ 2026-03-29 06:28:45,126 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=290 eval_loss=11.4031 duration_sec=13.90
88
+ 2026-03-29 06:29:27,812 [INFO] new_opacus_codex.train_steps: epoch=2 step=295 loss=11.3798
89
+ 2026-03-29 06:30:10,390 [INFO] new_opacus_codex.train_steps: epoch=2 step=300 loss=11.3804
90
+ 2026-03-29 06:30:24,281 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=300 eval_loss=11.2945 duration_sec=13.88
91
+ 2026-03-29 06:31:07,019 [INFO] new_opacus_codex.train_steps: epoch=2 step=305 loss=11.3463
92
+ 2026-03-29 06:31:50,841 [INFO] new_opacus_codex.train_steps: epoch=2 step=310 loss=11.2813
93
+ 2026-03-29 06:32:04,766 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=310 eval_loss=11.1933 duration_sec=13.91
94
+ 2026-03-29 06:32:47,650 [INFO] new_opacus_codex.train_steps: epoch=2 step=315 loss=11.2382
95
+ 2026-03-29 06:33:30,899 [INFO] new_opacus_codex.train_steps: epoch=2 step=320 loss=11.2982
96
+ 2026-03-29 06:33:44,808 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=320 eval_loss=11.0899 duration_sec=13.90
97
+ 2026-03-29 06:34:28,157 [INFO] new_opacus_codex.train_steps: epoch=2 step=325 loss=11.2292
98
+ 2026-03-29 06:35:10,145 [INFO] new_opacus_codex.train_steps: epoch=2 step=330 loss=11.1752
99
+ 2026-03-29 06:35:24,093 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=330 eval_loss=10.9892 duration_sec=13.94
100
+ 2026-03-29 06:36:05,256 [INFO] new_opacus_codex.train_steps: epoch=2 step=335 loss=11.1363
101
+ 2026-03-29 06:36:47,776 [INFO] new_opacus_codex.train_steps: epoch=2 step=340 loss=10.9588
102
+ 2026-03-29 06:37:01,684 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=340 eval_loss=10.8930 duration_sec=13.90
103
+ 2026-03-29 06:37:44,098 [INFO] new_opacus_codex.train_steps: epoch=2 step=345 loss=10.8470
104
+ 2026-03-29 06:38:28,909 [INFO] new_opacus_codex.train_steps: epoch=2 step=350 loss=10.8089
105
+ 2026-03-29 06:38:42,965 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=350 eval_loss=10.7921 duration_sec=14.04
106
+ 2026-03-29 06:39:24,953 [INFO] new_opacus_codex.train_steps: epoch=2 step=355 loss=10.9244
107
+ 2026-03-29 06:40:07,671 [INFO] new_opacus_codex.train_steps: epoch=2 step=360 loss=10.9922
108
+ 2026-03-29 06:40:21,563 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=360 eval_loss=10.6953 duration_sec=13.88
109
+ 2026-03-29 06:41:03,751 [INFO] new_opacus_codex.train_steps: epoch=2 step=365 loss=10.9054
110
+ 2026-03-29 06:42:11,453 [INFO] new_opacus_codex.train_steps: epoch=3 step=370 loss=10.6212
111
+ 2026-03-29 06:42:25,375 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=370 eval_loss=10.6050 duration_sec=13.91
112
+ 2026-03-29 06:43:08,088 [INFO] new_opacus_codex.train_steps: epoch=3 step=375 loss=10.5113
113
+ 2026-03-29 06:43:51,046 [INFO] new_opacus_codex.train_steps: epoch=3 step=380 loss=10.5553
114
+ 2026-03-29 06:44:04,954 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=380 eval_loss=10.5155 duration_sec=13.89
115
+ 2026-03-29 06:44:47,549 [INFO] new_opacus_codex.train_steps: epoch=3 step=385 loss=10.6349
116
+ 2026-03-29 06:45:29,847 [INFO] new_opacus_codex.train_steps: epoch=3 step=390 loss=10.5961
117
+ 2026-03-29 06:45:43,728 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=390 eval_loss=10.4222 duration_sec=13.87
118
+ 2026-03-29 06:46:26,259 [INFO] new_opacus_codex.train_steps: epoch=3 step=395 loss=10.5591
119
+ 2026-03-29 06:47:08,775 [INFO] new_opacus_codex.train_steps: epoch=3 step=400 loss=10.5074
120
+ 2026-03-29 06:47:22,711 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=400 eval_loss=10.3288 duration_sec=13.92
121
+ 2026-03-29 06:48:04,308 [INFO] new_opacus_codex.train_steps: epoch=3 step=405 loss=10.4153
122
+ 2026-03-29 06:48:48,031 [INFO] new_opacus_codex.train_steps: epoch=3 step=410 loss=10.4624
123
+ 2026-03-29 06:49:01,933 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=410 eval_loss=10.2313 duration_sec=13.89
124
+ 2026-03-29 06:49:44,102 [INFO] new_opacus_codex.train_steps: epoch=3 step=415 loss=10.3840
125
+ 2026-03-29 06:50:27,483 [INFO] new_opacus_codex.train_steps: epoch=3 step=420 loss=10.3176
126
+ 2026-03-29 06:50:41,458 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=420 eval_loss=10.1367 duration_sec=13.96
127
+ 2026-03-29 06:51:24,949 [INFO] new_opacus_codex.train_steps: epoch=3 step=425 loss=10.2497
128
+ 2026-03-29 06:52:08,419 [INFO] new_opacus_codex.train_steps: epoch=3 step=430 loss=10.1953
129
+ 2026-03-29 06:52:22,305 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=430 eval_loss=10.0555 duration_sec=13.88
130
+ 2026-03-29 06:53:04,853 [INFO] new_opacus_codex.train_steps: epoch=3 step=435 loss=10.1189
131
+ 2026-03-29 06:53:46,381 [INFO] new_opacus_codex.train_steps: epoch=3 step=440 loss=9.9982
132
+ 2026-03-29 06:54:00,328 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=440 eval_loss=9.9816 duration_sec=13.94
133
+ 2026-03-29 06:54:43,276 [INFO] new_opacus_codex.train_steps: epoch=3 step=445 loss=10.0861
134
+ 2026-03-29 06:55:26,030 [INFO] new_opacus_codex.train_steps: epoch=3 step=450 loss=10.1771
135
+ 2026-03-29 06:55:40,120 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=450 eval_loss=9.9103 duration_sec=14.08
136
+ 2026-03-29 06:56:21,988 [INFO] new_opacus_codex.train_steps: epoch=3 step=455 loss=10.2421
137
+ 2026-03-29 06:57:04,733 [INFO] new_opacus_codex.train_steps: epoch=3 step=460 loss=10.1535
138
+ 2026-03-29 06:57:18,720 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=460 eval_loss=9.8526 duration_sec=13.97
139
+ 2026-03-29 06:58:00,695 [INFO] new_opacus_codex.train_steps: epoch=3 step=465 loss=10.1080
140
+ 2026-03-29 06:58:42,642 [INFO] new_opacus_codex.train_steps: epoch=3 step=470 loss=10.0921
141
+ 2026-03-29 06:58:56,537 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=470 eval_loss=9.7915 duration_sec=13.88
142
+ 2026-03-29 06:59:39,040 [INFO] new_opacus_codex.train_steps: epoch=3 step=475 loss=10.1019
143
+ 2026-03-29 07:00:21,878 [INFO] new_opacus_codex.train_steps: epoch=3 step=480 loss=10.0799
144
+ 2026-03-29 07:00:35,782 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=480 eval_loss=9.7250 duration_sec=13.89
145
+ 2026-03-29 07:01:18,237 [INFO] new_opacus_codex.train_steps: epoch=3 step=485 loss=9.9324
146
+ 2026-03-29 07:02:01,752 [INFO] new_opacus_codex.train_steps: epoch=3 step=490 loss=9.9871
147
+ 2026-03-29 07:02:15,672 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=490 eval_loss=9.6696 duration_sec=13.85
148
+ 2026-03-29 07:02:58,217 [INFO] new_opacus_codex.train_steps: epoch=3 step=495 loss=10.0964
149
+ 2026-03-29 07:03:41,552 [INFO] new_opacus_codex.train_steps: epoch=3 step=500 loss=10.0276
150
+ 2026-03-29 07:03:55,403 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=500 eval_loss=9.6186 duration_sec=13.83
151
+ 2026-03-29 07:04:37,604 [INFO] new_opacus_codex.train_steps: epoch=3 step=505 loss=9.8982
152
+ 2026-03-29 07:05:20,809 [INFO] new_opacus_codex.train_steps: epoch=3 step=510 loss=9.9103
153
+ 2026-03-29 07:05:34,654 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=510 eval_loss=9.5682 duration_sec=13.83
154
+ 2026-03-29 07:06:18,058 [INFO] new_opacus_codex.train_steps: epoch=3 step=515 loss=9.8148
155
+ 2026-03-29 07:07:00,377 [INFO] new_opacus_codex.train_steps: epoch=3 step=520 loss=9.8351
156
+ 2026-03-29 07:07:14,259 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=520 eval_loss=9.5232 duration_sec=13.87
157
+ 2026-03-29 07:07:55,631 [INFO] new_opacus_codex.train_steps: epoch=3 step=525 loss=9.8304
158
+ 2026-03-29 07:08:37,647 [INFO] new_opacus_codex.train_steps: epoch=3 step=530 loss=9.6276
159
+ 2026-03-29 07:08:51,572 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=530 eval_loss=9.4853 duration_sec=13.91
160
+ 2026-03-29 07:09:33,425 [INFO] new_opacus_codex.train_steps: epoch=3 step=535 loss=9.8235
161
+ 2026-03-29 07:10:16,837 [INFO] new_opacus_codex.train_steps: epoch=3 step=540 loss=9.7965
162
+ 2026-03-29 07:10:30,721 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=540 eval_loss=9.4484 duration_sec=13.84
163
+ 2026-03-29 07:11:13,813 [INFO] new_opacus_codex.train_steps: epoch=3 step=545 loss=9.6728
164
+ 2026-03-29 07:11:56,493 [INFO] new_opacus_codex.train_steps: epoch=3 step=550 loss=9.6830
165
+ 2026-03-29 07:12:10,341 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=550 eval_loss=9.4117 duration_sec=13.63
166
+ 2026-03-29 07:13:15,474 [INFO] new_opacus_codex.train_steps: epoch=4 step=555 loss=9.6268
167
+ 2026-03-29 07:14:01,405 [INFO] new_opacus_codex.train_steps: epoch=4 step=560 loss=9.5461
168
+ 2026-03-29 07:14:15,255 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=560 eval_loss=9.3772 duration_sec=13.84
169
+ 2026-03-29 07:14:57,788 [INFO] new_opacus_codex.train_steps: epoch=4 step=565 loss=9.5851
170
+ 2026-03-29 07:15:40,843 [INFO] new_opacus_codex.train_steps: epoch=4 step=570 loss=9.6353
171
+ 2026-03-29 07:15:54,712 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=570 eval_loss=9.3416 duration_sec=13.86
172
+ 2026-03-29 07:16:37,670 [INFO] new_opacus_codex.train_steps: epoch=4 step=575 loss=9.5442
173
+ 2026-03-29 07:17:20,424 [INFO] new_opacus_codex.train_steps: epoch=4 step=580 loss=9.4937
174
+ 2026-03-29 07:17:34,304 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=580 eval_loss=9.3096 duration_sec=13.87
175
+ 2026-03-29 07:18:18,214 [INFO] new_opacus_codex.train_steps: epoch=4 step=585 loss=9.5924
176
+ 2026-03-29 07:19:00,174 [INFO] new_opacus_codex.train_steps: epoch=4 step=590 loss=9.6653
177
+ 2026-03-29 07:19:14,064 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=590 eval_loss=9.2792 duration_sec=13.88
178
+ 2026-03-29 07:19:56,370 [INFO] new_opacus_codex.train_steps: epoch=4 step=595 loss=9.5346
179
+ 2026-03-29 07:20:40,399 [INFO] new_opacus_codex.train_steps: epoch=4 step=600 loss=9.3879
180
+ 2026-03-29 07:20:54,294 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=600 eval_loss=9.2507 duration_sec=13.88
181
+ 2026-03-29 07:21:36,152 [INFO] new_opacus_codex.train_steps: epoch=4 step=605 loss=9.5244
182
+ 2026-03-29 07:22:19,272 [INFO] new_opacus_codex.train_steps: epoch=4 step=610 loss=9.5839
183
+ 2026-03-29 07:22:33,134 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=610 eval_loss=9.2214 duration_sec=13.85
184
+ 2026-03-29 07:23:15,865 [INFO] new_opacus_codex.train_steps: epoch=4 step=615 loss=9.5453
185
+ 2026-03-29 07:23:58,035 [INFO] new_opacus_codex.train_steps: epoch=4 step=620 loss=9.6512
186
+ 2026-03-29 07:24:11,990 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=620 eval_loss=9.1933 duration_sec=13.95
187
+ 2026-03-29 07:24:55,658 [INFO] new_opacus_codex.train_steps: epoch=4 step=625 loss=9.5778
188
+ 2026-03-29 07:25:38,972 [INFO] new_opacus_codex.train_steps: epoch=4 step=630 loss=9.5521
189
+ 2026-03-29 07:25:52,852 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=630 eval_loss=9.1668 duration_sec=13.87
190
+ 2026-03-29 07:26:35,908 [INFO] new_opacus_codex.train_steps: epoch=4 step=635 loss=9.6073
191
+ 2026-03-29 07:27:19,282 [INFO] new_opacus_codex.train_steps: epoch=4 step=640 loss=9.5242
192
+ 2026-03-29 07:27:33,144 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=640 eval_loss=9.1452 duration_sec=13.85
193
+ 2026-03-29 07:28:16,726 [INFO] new_opacus_codex.train_steps: epoch=4 step=645 loss=9.4743
194
+ 2026-03-29 07:28:59,182 [INFO] new_opacus_codex.train_steps: epoch=4 step=650 loss=9.4579
195
+ 2026-03-29 07:29:13,043 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=650 eval_loss=9.1234 duration_sec=13.85
196
+ 2026-03-29 07:29:55,311 [INFO] new_opacus_codex.train_steps: epoch=4 step=655 loss=9.4626
197
+ 2026-03-29 07:30:38,816 [INFO] new_opacus_codex.train_steps: epoch=4 step=660 loss=9.3040
198
+ 2026-03-29 07:30:52,719 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=660 eval_loss=9.1043 duration_sec=13.90
199
+ 2026-03-29 07:31:35,043 [INFO] new_opacus_codex.train_steps: epoch=4 step=665 loss=9.3262
200
+ 2026-03-29 07:32:17,323 [INFO] new_opacus_codex.train_steps: epoch=4 step=670 loss=9.4976
201
+ 2026-03-29 07:32:31,268 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=670 eval_loss=9.0864 duration_sec=13.93
202
+ 2026-03-29 07:33:14,351 [INFO] new_opacus_codex.train_steps: epoch=4 step=675 loss=9.4661
203
+ 2026-03-29 07:33:56,428 [INFO] new_opacus_codex.train_steps: epoch=4 step=680 loss=9.3957
204
+ 2026-03-29 07:34:10,355 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=680 eval_loss=9.0684 duration_sec=13.91
205
+ 2026-03-29 07:34:53,169 [INFO] new_opacus_codex.train_steps: epoch=4 step=685 loss=9.2888
206
+ 2026-03-29 07:35:37,770 [INFO] new_opacus_codex.train_steps: epoch=4 step=690 loss=9.3255
207
+ 2026-03-29 07:35:51,704 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=690 eval_loss=9.0510 duration_sec=13.92
208
+ 2026-03-29 07:36:34,504 [INFO] new_opacus_codex.train_steps: epoch=4 step=695 loss=9.4954
209
+ 2026-03-29 07:37:17,542 [INFO] new_opacus_codex.train_steps: epoch=4 step=700 loss=9.4823
210
+ 2026-03-29 07:37:31,564 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=700 eval_loss=9.0324 duration_sec=14.01
211
+ 2026-03-29 07:38:13,788 [INFO] new_opacus_codex.train_steps: epoch=4 step=705 loss=9.3420
212
+ 2026-03-29 07:38:56,292 [INFO] new_opacus_codex.train_steps: epoch=4 step=710 loss=9.4012
213
+ 2026-03-29 07:39:10,259 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=710 eval_loss=9.0158 duration_sec=13.95
214
+ 2026-03-29 07:39:52,818 [INFO] new_opacus_codex.train_steps: epoch=4 step=715 loss=9.3514
215
+ 2026-03-29 07:40:36,255 [INFO] new_opacus_codex.train_steps: epoch=4 step=720 loss=9.3579
216
+ 2026-03-29 07:40:50,236 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=720 eval_loss=9.0013 duration_sec=13.97
217
+ 2026-03-29 07:41:33,743 [INFO] new_opacus_codex.train_steps: epoch=4 step=725 loss=9.3657
218
+ 2026-03-29 07:42:16,995 [INFO] new_opacus_codex.train_steps: epoch=4 step=730 loss=9.3126
219
+ 2026-03-29 07:42:30,982 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=730 eval_loss=8.9876 duration_sec=13.95
220
+ 2026-03-29 07:43:14,800 [INFO] new_opacus_codex.train_steps: epoch=4 step=735 loss=9.3190
221
+ 2026-03-29 07:44:21,737 [INFO] new_opacus_codex.train_steps: epoch=5 step=740 loss=9.4568
222
+ 2026-03-29 07:44:35,630 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=740 eval_loss=8.9751 duration_sec=13.86
223
+ 2026-03-29 07:45:17,905 [INFO] new_opacus_codex.train_steps: epoch=5 step=745 loss=9.5412
224
+ 2026-03-29 07:46:00,572 [INFO] new_opacus_codex.train_steps: epoch=5 step=750 loss=9.3034
225
+ 2026-03-29 07:46:14,504 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=750 eval_loss=8.9632 duration_sec=13.80
226
+ 2026-03-29 07:46:59,127 [INFO] new_opacus_codex.train_steps: epoch=5 step=755 loss=9.1359
227
+ 2026-03-29 07:47:41,707 [INFO] new_opacus_codex.train_steps: epoch=5 step=760 loss=9.3536
228
+ 2026-03-29 07:47:55,791 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=760 eval_loss=8.9517 duration_sec=14.07
229
+ 2026-03-29 07:48:38,874 [INFO] new_opacus_codex.train_steps: epoch=5 step=765 loss=9.3280
230
+ 2026-03-29 07:49:21,316 [INFO] new_opacus_codex.train_steps: epoch=5 step=770 loss=9.2614
231
+ 2026-03-29 07:49:35,224 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=770 eval_loss=8.9401 duration_sec=13.80
232
+ 2026-03-29 07:50:17,896 [INFO] new_opacus_codex.train_steps: epoch=5 step=775 loss=9.3448
233
+ 2026-03-29 07:51:00,278 [INFO] new_opacus_codex.train_steps: epoch=5 step=780 loss=9.4861
234
+ 2026-03-29 07:51:14,170 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=780 eval_loss=8.9306 duration_sec=13.78
235
+ 2026-03-29 07:51:56,598 [INFO] new_opacus_codex.train_steps: epoch=5 step=785 loss=9.4331
236
+ 2026-03-29 07:52:39,325 [INFO] new_opacus_codex.train_steps: epoch=5 step=790 loss=9.3122
237
+ 2026-03-29 07:52:53,358 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=790 eval_loss=8.9200 duration_sec=13.93
238
+ 2026-03-29 07:53:35,636 [INFO] new_opacus_codex.train_steps: epoch=5 step=795 loss=9.3591
239
+ 2026-03-29 07:54:18,980 [INFO] new_opacus_codex.train_steps: epoch=5 step=800 loss=9.4804
240
+ 2026-03-29 07:54:32,896 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=800 eval_loss=8.9081 duration_sec=13.84
241
+ 2026-03-29 07:55:15,237 [INFO] new_opacus_codex.train_steps: epoch=5 step=805 loss=9.4989
242
+ 2026-03-29 07:55:58,101 [INFO] new_opacus_codex.train_steps: epoch=5 step=810 loss=9.3818
243
+ 2026-03-29 07:56:12,125 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=810 eval_loss=8.8967 duration_sec=14.01
244
+ 2026-03-29 07:56:54,534 [INFO] new_opacus_codex.train_steps: epoch=5 step=815 loss=9.2909
245
+ 2026-03-29 07:57:38,502 [INFO] new_opacus_codex.train_steps: epoch=5 step=820 loss=9.0823
246
+ 2026-03-29 07:57:52,405 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=820 eval_loss=8.8871 duration_sec=13.86
247
+ 2026-03-29 07:58:34,228 [INFO] new_opacus_codex.train_steps: epoch=5 step=825 loss=9.0702
248
+ 2026-03-29 07:59:16,970 [INFO] new_opacus_codex.train_steps: epoch=5 step=830 loss=9.1519
249
+ 2026-03-29 07:59:30,879 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=830 eval_loss=8.8787 duration_sec=13.88
250
+ 2026-03-29 08:00:16,986 [INFO] new_opacus_codex.train_steps: epoch=5 step=835 loss=8.9837
251
+ 2026-03-29 08:01:00,171 [INFO] new_opacus_codex.train_steps: epoch=5 step=840 loss=9.1147
252
+ 2026-03-29 08:01:14,165 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=840 eval_loss=8.8702 duration_sec=13.99
253
+ 2026-03-29 08:01:57,270 [INFO] new_opacus_codex.train_steps: epoch=5 step=845 loss=9.1954
254
+ 2026-03-29 08:02:38,705 [INFO] new_opacus_codex.train_steps: epoch=5 step=850 loss=9.1040
255
+ 2026-03-29 08:02:52,686 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=850 eval_loss=8.8628 duration_sec=13.94
256
+ 2026-03-29 08:03:36,754 [INFO] new_opacus_codex.train_steps: epoch=5 step=855 loss=9.1987
257
+ 2026-03-29 08:04:19,702 [INFO] new_opacus_codex.train_steps: epoch=5 step=860 loss=9.0988
258
+ 2026-03-29 08:04:33,661 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=860 eval_loss=8.8564 duration_sec=13.72
259
+ 2026-03-29 08:05:16,838 [INFO] new_opacus_codex.train_steps: epoch=5 step=865 loss=9.0361
260
+ 2026-03-29 08:05:59,205 [INFO] new_opacus_codex.train_steps: epoch=5 step=870 loss=9.1400
261
+ 2026-03-29 08:06:13,128 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=870 eval_loss=8.8505 duration_sec=13.92
262
+ 2026-03-29 08:06:57,031 [INFO] new_opacus_codex.train_steps: epoch=5 step=875 loss=9.1541
263
+ 2026-03-29 08:07:41,240 [INFO] new_opacus_codex.train_steps: epoch=5 step=880 loss=9.1480
264
+ 2026-03-29 08:07:55,200 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=880 eval_loss=8.8442 duration_sec=13.95
265
+ 2026-03-29 08:08:39,920 [INFO] new_opacus_codex.train_steps: epoch=5 step=885 loss=9.2041
266
+ 2026-03-29 08:09:23,336 [INFO] new_opacus_codex.train_steps: epoch=5 step=890 loss=9.1408
267
+ 2026-03-29 08:09:37,248 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=890 eval_loss=8.8369 duration_sec=13.91
268
+ 2026-03-29 08:10:21,428 [INFO] new_opacus_codex.train_steps: epoch=5 step=895 loss=9.0693
269
+ 2026-03-29 08:11:05,799 [INFO] new_opacus_codex.train_steps: epoch=5 step=900 loss=9.1123
270
+ 2026-03-29 08:11:19,804 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=900 eval_loss=8.8294 duration_sec=13.99
271
+ 2026-03-29 08:12:03,122 [INFO] new_opacus_codex.train_steps: epoch=5 step=905 loss=9.3328
272
+ 2026-03-29 08:12:46,987 [INFO] new_opacus_codex.train_steps: epoch=5 step=910 loss=9.2737
273
+ 2026-03-29 08:13:00,974 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=910 eval_loss=8.8220 duration_sec=13.97
274
+ 2026-03-29 08:13:44,920 [INFO] new_opacus_codex.train_steps: epoch=5 step=915 loss=9.1627
275
+ 2026-03-29 08:14:21,544 [INFO] new_opacus_codex.train_steps: epoch=5 step=920 loss=9.3556
276
+ 2026-03-29 08:14:35,451 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=920 eval_loss=8.8145 duration_sec=13.90
llama3.1-8b/dp8/adapter/README.md ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.1-8B
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:meta-llama/Llama-3.1-8B
7
+ - lora
8
+ - transformers
9
+ ---
10
+
11
+ # Model Card for Model ID
12
+
13
+ <!-- Provide a quick summary of what the model is/does. -->
14
+
15
+
16
+
17
+ ## Model Details
18
+
19
+ ### Model Description
20
+
21
+ <!-- Provide a longer summary of what this model is. -->
22
+
23
+
24
+
25
+ - **Developed by:** [More Information Needed]
26
+ - **Funded by [optional]:** [More Information Needed]
27
+ - **Shared by [optional]:** [More Information Needed]
28
+ - **Model type:** [More Information Needed]
29
+ - **Language(s) (NLP):** [More Information Needed]
30
+ - **License:** [More Information Needed]
31
+ - **Finetuned from model [optional]:** [More Information Needed]
32
+
33
+ ### Model Sources [optional]
34
+
35
+ <!-- Provide the basic links for the model. -->
36
+
37
+ - **Repository:** [More Information Needed]
38
+ - **Paper [optional]:** [More Information Needed]
39
+ - **Demo [optional]:** [More Information Needed]
40
+
41
+ ## Uses
42
+
43
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
44
+
45
+ ### Direct Use
46
+
47
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
48
+
49
+ [More Information Needed]
50
+
51
+ ### Downstream Use [optional]
52
+
53
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
54
+
55
+ [More Information Needed]
56
+
57
+ ### Out-of-Scope Use
58
+
59
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
60
+
61
+ [More Information Needed]
62
+
63
+ ## Bias, Risks, and Limitations
64
+
65
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
66
+
67
+ [More Information Needed]
68
+
69
+ ### Recommendations
70
+
71
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
72
+
73
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
74
+
75
+ ## How to Get Started with the Model
76
+
77
+ Use the code below to get started with the model.
78
+
79
+ [More Information Needed]
80
+
81
+ ## Training Details
82
+
83
+ ### Training Data
84
+
85
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
86
+
87
+ [More Information Needed]
88
+
89
+ ### Training Procedure
90
+
91
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
92
+
93
+ #### Preprocessing [optional]
94
+
95
+ [More Information Needed]
96
+
97
+
98
+ #### Training Hyperparameters
99
+
100
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
101
+
102
+ #### Speeds, Sizes, Times [optional]
103
+
104
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
105
+
106
+ [More Information Needed]
107
+
108
+ ## Evaluation
109
+
110
+ <!-- This section describes the evaluation protocols and provides the results. -->
111
+
112
+ ### Testing Data, Factors & Metrics
113
+
114
+ #### Testing Data
115
+
116
+ <!-- This should link to a Dataset Card if possible. -->
117
+
118
+ [More Information Needed]
119
+
120
+ #### Factors
121
+
122
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
123
+
124
+ [More Information Needed]
125
+
126
+ #### Metrics
127
+
128
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
129
+
130
+ [More Information Needed]
131
+
132
+ ### Results
133
+
134
+ [More Information Needed]
135
+
136
+ #### Summary
137
+
138
+
139
+
140
+ ## Model Examination [optional]
141
+
142
+ <!-- Relevant interpretability work for the model goes here -->
143
+
144
+ [More Information Needed]
145
+
146
+ ## Environmental Impact
147
+
148
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
149
+
150
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
151
+
152
+ - **Hardware Type:** [More Information Needed]
153
+ - **Hours used:** [More Information Needed]
154
+ - **Cloud Provider:** [More Information Needed]
155
+ - **Compute Region:** [More Information Needed]
156
+ - **Carbon Emitted:** [More Information Needed]
157
+
158
+ ## Technical Specifications [optional]
159
+
160
+ ### Model Architecture and Objective
161
+
162
+ [More Information Needed]
163
+
164
+ ### Compute Infrastructure
165
+
166
+ [More Information Needed]
167
+
168
+ #### Hardware
169
+
170
+ [More Information Needed]
171
+
172
+ #### Software
173
+
174
+ [More Information Needed]
175
+
176
+ ## Citation [optional]
177
+
178
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
179
+
180
+ **BibTeX:**
181
+
182
+ [More Information Needed]
183
+
184
+ **APA:**
185
+
186
+ [More Information Needed]
187
+
188
+ ## Glossary [optional]
189
+
190
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
191
+
192
+ [More Information Needed]
193
+
194
+ ## More Information [optional]
195
+
196
+ [More Information Needed]
197
+
198
+ ## Model Card Authors [optional]
199
+
200
+ [More Information Needed]
201
+
202
+ ## Model Card Contact
203
+
204
+ [More Information Needed]
205
+ ### Framework versions
206
+
207
+ - PEFT 0.18.1
llama3.1-8b/dp8/adapter/adapter_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "meta-llama/Llama-3.1-8B",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": true,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": [
25
+ "lm_head",
26
+ "embed_tokens"
27
+ ],
28
+ "peft_type": "LORA",
29
+ "peft_version": "0.18.1",
30
+ "qalora_group_size": 16,
31
+ "r": 16,
32
+ "rank_pattern": {},
33
+ "revision": null,
34
+ "target_modules": [
35
+ "v_proj",
36
+ "o_proj",
37
+ "q_proj",
38
+ "k_proj"
39
+ ],
40
+ "target_parameters": null,
41
+ "task_type": "CAUSAL_LM",
42
+ "trainable_token_indices": null,
43
+ "use_dora": false,
44
+ "use_qalora": false,
45
+ "use_rslora": false
46
+ }
llama3.1-8b/dp8/audit_results.json ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "delta": 1e-05,
3
+ "num_canaries": 500,
4
+ "num_members": 250,
5
+ "paper_guess_fraction": 0.2,
6
+ "paper_guess_steps": 20,
7
+ "loss": {
8
+ "auc": 0.523296,
9
+ "empirical_epsilon": {
10
+ "0.05": 0.054578399285674095,
11
+ "0.01": 0.0
12
+ },
13
+ "empirical_epsilon_details": {
14
+ "0.05": {
15
+ "epsilon": 0.054578399285674095,
16
+ "num_guesses": 50,
17
+ "correct_guesses": 32,
18
+ "candidate_num_guesses": [
19
+ 5,
20
+ 10,
21
+ 15,
22
+ 20,
23
+ 25,
24
+ 30,
25
+ 35,
26
+ 40,
27
+ 45,
28
+ 50,
29
+ 55,
30
+ 60,
31
+ 65,
32
+ 70,
33
+ 75,
34
+ 80,
35
+ 85,
36
+ 90,
37
+ 95,
38
+ 100
39
+ ],
40
+ "direction": "lower"
41
+ },
42
+ "0.01": {
43
+ "epsilon": 0.0,
44
+ "num_guesses": 0,
45
+ "correct_guesses": 0,
46
+ "candidate_num_guesses": [
47
+ 5,
48
+ 10,
49
+ 15,
50
+ 20,
51
+ 25,
52
+ 30,
53
+ 35,
54
+ 40,
55
+ 45,
56
+ 50,
57
+ 55,
58
+ 60,
59
+ 65,
60
+ 70,
61
+ 75,
62
+ 80,
63
+ 85,
64
+ 90,
65
+ 95,
66
+ 100
67
+ ],
68
+ "direction": "lower"
69
+ }
70
+ }
71
+ },
72
+ "embedding": {
73
+ "auc": 0.500064,
74
+ "empirical_epsilon": {
75
+ "0.05": 0.0,
76
+ "0.01": 0.0
77
+ },
78
+ "empirical_epsilon_details": {
79
+ "0.05": {
80
+ "epsilon": 0.0,
81
+ "num_guesses": 0,
82
+ "correct_guesses": 0,
83
+ "candidate_num_guesses": [
84
+ 5,
85
+ 10,
86
+ 15,
87
+ 20,
88
+ 25,
89
+ 30,
90
+ 35,
91
+ 40,
92
+ 45,
93
+ 50,
94
+ 55,
95
+ 60,
96
+ 65,
97
+ 70,
98
+ 75,
99
+ 80,
100
+ 85,
101
+ 90,
102
+ 95,
103
+ 100
104
+ ],
105
+ "direction": "lower"
106
+ },
107
+ "0.01": {
108
+ "epsilon": 0.0,
109
+ "num_guesses": 0,
110
+ "correct_guesses": 0,
111
+ "candidate_num_guesses": [
112
+ 5,
113
+ 10,
114
+ 15,
115
+ 20,
116
+ 25,
117
+ 30,
118
+ 35,
119
+ 40,
120
+ 45,
121
+ 50,
122
+ 55,
123
+ 60,
124
+ 65,
125
+ 70,
126
+ 75,
127
+ 80,
128
+ 85,
129
+ 90,
130
+ 95,
131
+ 100
132
+ ],
133
+ "direction": "lower"
134
+ }
135
+ }
136
+ }
137
+ }
llama3.1-8b/dp8/canary_meta.json ADDED
The diff for this file is too large to render. See raw diff
 
llama3.1-8b/dp8/metrics.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
llama3.1-8b/dp8/resolved_config.yaml ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ name: meta-llama/Llama-3.1-8B
3
+ tokenizer_name: meta-llama/Llama-3.1-8B
4
+ max_length: 1024
5
+ dtype: bfloat16
6
+ trust_remote_code: true
7
+ use_fast_tokenizer: true
8
+ cache_dir: null
9
+ local_files_only: false
10
+ low_cpu_mem_usage: true
11
+ tie_word_embeddings: true
12
+ gradient_checkpointing: false
13
+ use_chat_template: false
14
+ dataset:
15
+ name: melihcatal/codedp-cpt
16
+ split: train
17
+ mode: cpt
18
+ text_column: text
19
+ validation_ratio: 0.05
20
+ max_samples: -1
21
+ lora:
22
+ enabled: true
23
+ r: 16
24
+ alpha: 32
25
+ dropout: 0.05
26
+ target_modules:
27
+ - q_proj
28
+ - k_proj
29
+ - v_proj
30
+ - o_proj
31
+ modules_to_save:
32
+ - lm_head
33
+ bias: none
34
+ training:
35
+ seed: 42
36
+ epochs: 5
37
+ warmup_steps: null
38
+ warmup_ratio: 0.05
39
+ mixed_precision: false
40
+ mixed_precision_dtype: bfloat16
41
+ batch_size: 4
42
+ eval_batch_size: 8
43
+ eval_every_steps: 10
44
+ eval_every_epochs: 1
45
+ learning_rate: 0.0005
46
+ optimizer: adamw
47
+ lr_scheduler: cosine
48
+ adam_beta1: 0.9
49
+ adam_beta2: 0.999
50
+ adam_epsilon: 1.0e-08
51
+ sgd_momentum: 0.9
52
+ weight_decay: 0.01
53
+ max_grad_norm: 1.0
54
+ log_every: 5
55
+ gradient_accumulation_steps: 8
56
+ num_workers: 4
57
+ output_dir: /scratch/mcatal/runs/cpt/llama3.1-8b/dp8
58
+ min_lr_ratio: 0.15
59
+ distributed:
60
+ strategy: dpddp
61
+ backend: nccl
62
+ devices: null
63
+ dp:
64
+ module_validator: auto
65
+ target_delta: 1.0e-05
66
+ noise_multiplier: null
67
+ max_grad_norm: 1.0
68
+ grad_sample_mode: hooks
69
+ clipping: per_layer
70
+ secure_mode: false
71
+ enabled: true
72
+ target_epsilon: 8.0
73
+ audit:
74
+ enabled: true
75
+ run_every_epoch: true
76
+ epoch_device: cuda
77
+ q_canary: auto
78
+ num_canaries: 500
79
+ prefix_length: 49
80
+ num_digits: 12
81
+ batch_size: 32
82
+ delta: 1.0e-05
83
+ p_values:
84
+ - 0.05
85
+ - 0.01
86
+ paper_guess_fraction: 0.2
87
+ paper_guess_steps: 20
88
+ enable_holdout_empirical_epsilon: false
89
+ holdout_seed: 42
90
+ tie_seed: 42
91
+ tracking:
92
+ enabled: true
93
+ tensorboard: true
94
+ wandb: false
95
+ wandb_project: codedp-finetune-h200-audit
96
+ wandb_run_name: llama3.1-8b-cpt-dp8
97
+ wandb_mode: online
98
+ codecarbon: true
99
+ codecarbon_output_file: codecarbon.csv
100
+ codecarbon_measure_power_secs: 15
101
+ codecarbon_country_iso_code: null
102
+ codecarbon_project_name: codedp-llama3.1-8b-cpt-dp8
llama3.1-8b/dp8/summary.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audit/delta": 1e-05,
3
+ "audit/embedding/auc": 0.500064,
4
+ "audit/embedding/empirical_epsilon/0.01": 0.0,
5
+ "audit/embedding/empirical_epsilon/0.05": 0.0,
6
+ "audit/embedding/empirical_epsilon_details/0.01/correct_guesses": 0.0,
7
+ "audit/embedding/empirical_epsilon_details/0.01/epsilon": 0.0,
8
+ "audit/embedding/empirical_epsilon_details/0.01/num_guesses": 0.0,
9
+ "audit/embedding/empirical_epsilon_details/0.05/correct_guesses": 0.0,
10
+ "audit/embedding/empirical_epsilon_details/0.05/epsilon": 0.0,
11
+ "audit/embedding/empirical_epsilon_details/0.05/num_guesses": 0.0,
12
+ "audit/loss/auc": 0.523296,
13
+ "audit/loss/empirical_epsilon/0.01": 0.0,
14
+ "audit/loss/empirical_epsilon/0.05": 0.054578399285674095,
15
+ "audit/loss/empirical_epsilon_details/0.01/correct_guesses": 0.0,
16
+ "audit/loss/empirical_epsilon_details/0.01/epsilon": 0.0,
17
+ "audit/loss/empirical_epsilon_details/0.01/num_guesses": 0.0,
18
+ "audit/loss/empirical_epsilon_details/0.05/correct_guesses": 32.0,
19
+ "audit/loss/empirical_epsilon_details/0.05/epsilon": 0.054578399285674095,
20
+ "audit/loss/empirical_epsilon_details/0.05/num_guesses": 50.0,
21
+ "audit/num_canaries": 500.0,
22
+ "audit/num_members": 250.0,
23
+ "audit/paper_guess_fraction": 0.2,
24
+ "audit/paper_guess_steps": 20.0,
25
+ "energy/codecarbon/cpu_count": 16.0,
26
+ "energy/codecarbon/cpu_energy": 0.20345157926762186,
27
+ "energy/codecarbon/cpu_power": 80.03159974853541,
28
+ "energy/codecarbon/cpu_utilization_percent": 3.5419628647214854,
29
+ "energy/codecarbon/duration": 9512.013140782248,
30
+ "energy/codecarbon/emissions": 0.3276752594920991,
31
+ "energy/codecarbon/emissions_rate": 3.444857094311707e-05,
32
+ "energy/codecarbon/energy_consumed": 9.404335433002297,
33
+ "energy/codecarbon/gpu_count": 8.0,
34
+ "energy/codecarbon/gpu_energy": 9.104283212865184,
35
+ "energy/codecarbon/gpu_power": 3457.6489583314255,
36
+ "energy/codecarbon/gpu_utilization_percent": 95.43501326259947,
37
+ "energy/codecarbon/latitude": 47.4843,
38
+ "energy/codecarbon/longitude": 8.212,
39
+ "energy/codecarbon/pue": 1.0,
40
+ "energy/codecarbon/ram_energy": 0.09660064086948418,
41
+ "energy/codecarbon/ram_power": 38.0,
42
+ "energy/codecarbon/ram_total_size": 128.0,
43
+ "energy/codecarbon/ram_used_gb": 509.469642048641,
44
+ "energy/codecarbon/ram_utilization_percent": 25.71827055702918,
45
+ "energy/codecarbon/water_consumed": 0.0,
46
+ "energy/codecarbon/wue": 0.0,
47
+ "eval/duration_sec": 13.953276729211211,
48
+ "eval/loss": 7.627163219146239,
49
+ "perf/audit_duration_sec": 7.110759837087244,
50
+ "perf/epoch_duration_sec": 1873.1702182018198,
51
+ "perf/epoch_samples": 51412.0,
52
+ "perf/epoch_samples_per_sec": 27.446517940774108,
53
+ "perf/epoch_tokens": 37419500.0,
54
+ "perf/epoch_tokens_per_sec": 19976.56146589895,
55
+ "perf/logical_batch_size": 3.0,
56
+ "perf/logical_token_count": 432.0,
57
+ "perf/physical_batches": 2.0,
58
+ "perf/samples_per_sec": 1.6736803066877797,
59
+ "perf/step_duration_sec": 1.7924570110626519,
60
+ "perf/tokens_per_sec": 241.00996416304028,
61
+ "privacy/epsilon": 7.992470385789166,
62
+ "system/cuda_epoch_peak_memory_gb": 62.821791648864746,
63
+ "system/cuda_max_memory_allocated_gb": 62.821791648864746,
64
+ "system/cuda_memory_allocated_gb": 20.329766750335693,
65
+ "train/epoch_canary_loss": 10.347639968544451,
66
+ "train/epoch_loss": 8.10246930775083,
67
+ "train/epoch_real_loss": 7.769937282630607,
68
+ "train/lr": 7.5e-05,
69
+ "train/step_canary_loss": 9.416666666666666,
70
+ "train/step_loss": 8.914817333221436,
71
+ "train/step_real_loss": 7.409266471862793
72
+ }
llama3.1-8b/dp8/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,516 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "bos_token": "<|begin_of_text|>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|end_of_text|>",
6
+ "extra_special_tokens": {
7
+ "extra_special_token_0": "865331112869",
8
+ "extra_special_token_1": "569765693871",
9
+ "extra_special_token_2": "485177821815",
10
+ "extra_special_token_3": "135441121756",
11
+ "extra_special_token_4": "367459894796",
12
+ "extra_special_token_5": "877482678543",
13
+ "extra_special_token_6": "457919547633",
14
+ "extra_special_token_7": "765474393376",
15
+ "extra_special_token_8": "114848338811",
16
+ "extra_special_token_9": "746285987371",
17
+ "extra_special_token_10": "649291669397",
18
+ "extra_special_token_11": "927914615679",
19
+ "extra_special_token_12": "445925149649",
20
+ "extra_special_token_13": "691587454538",
21
+ "extra_special_token_14": "143777992227",
22
+ "extra_special_token_15": "997981281989",
23
+ "extra_special_token_16": "425949483533",
24
+ "extra_special_token_17": "982993456429",
25
+ "extra_special_token_18": "718726519731",
26
+ "extra_special_token_19": "172599315861",
27
+ "extra_special_token_20": "643489267333",
28
+ "extra_special_token_21": "282322838685",
29
+ "extra_special_token_22": "781653545886",
30
+ "extra_special_token_23": "796415361892",
31
+ "extra_special_token_24": "841991688488",
32
+ "extra_special_token_25": "211411365397",
33
+ "extra_special_token_26": "698218415444",
34
+ "extra_special_token_27": "355977139358",
35
+ "extra_special_token_28": "682564697312",
36
+ "extra_special_token_29": "383837596997",
37
+ "extra_special_token_30": "689362171782",
38
+ "extra_special_token_31": "749966767285",
39
+ "extra_special_token_32": "753159165157",
40
+ "extra_special_token_33": "795693824762",
41
+ "extra_special_token_34": "669689115557",
42
+ "extra_special_token_35": "327491773134",
43
+ "extra_special_token_36": "983569279932",
44
+ "extra_special_token_37": "612128769512",
45
+ "extra_special_token_38": "374327157578",
46
+ "extra_special_token_39": "311632789559",
47
+ "extra_special_token_40": "523918658846",
48
+ "extra_special_token_41": "765981581453",
49
+ "extra_special_token_42": "794825141891",
50
+ "extra_special_token_43": "873898736873",
51
+ "extra_special_token_44": "447445629421",
52
+ "extra_special_token_45": "473822473819",
53
+ "extra_special_token_46": "181439694557",
54
+ "extra_special_token_47": "592538279337",
55
+ "extra_special_token_48": "668134915514",
56
+ "extra_special_token_49": "643692393748",
57
+ "extra_special_token_50": "696651276628",
58
+ "extra_special_token_51": "853859348234",
59
+ "extra_special_token_52": "778466723723",
60
+ "extra_special_token_53": "929826356991",
61
+ "extra_special_token_54": "272362973463",
62
+ "extra_special_token_55": "694235616268",
63
+ "extra_special_token_56": "281673864127",
64
+ "extra_special_token_57": "479676316326",
65
+ "extra_special_token_58": "646979124677",
66
+ "extra_special_token_59": "922327493433",
67
+ "extra_special_token_60": "883685933161",
68
+ "extra_special_token_61": "264259917554",
69
+ "extra_special_token_62": "836746273134",
70
+ "extra_special_token_63": "658481324922",
71
+ "extra_special_token_64": "481884157827",
72
+ "extra_special_token_65": "587787496812",
73
+ "extra_special_token_66": "579184949249",
74
+ "extra_special_token_67": "912193598348",
75
+ "extra_special_token_68": "529679678956",
76
+ "extra_special_token_69": "795838284624",
77
+ "extra_special_token_70": "159337222655",
78
+ "extra_special_token_71": "173781362446",
79
+ "extra_special_token_72": "773687856563",
80
+ "extra_special_token_73": "535787224917",
81
+ "extra_special_token_74": "351885857332",
82
+ "extra_special_token_75": "578827344666",
83
+ "extra_special_token_76": "198462689911",
84
+ "extra_special_token_77": "722618266242",
85
+ "extra_special_token_78": "952872416512",
86
+ "extra_special_token_79": "517778845323",
87
+ "extra_special_token_80": "749665846687",
88
+ "extra_special_token_81": "661436365453",
89
+ "extra_special_token_82": "259666844669",
90
+ "extra_special_token_83": "242851284913",
91
+ "extra_special_token_84": "514532995959",
92
+ "extra_special_token_85": "161588262349",
93
+ "extra_special_token_86": "742765629356",
94
+ "extra_special_token_87": "225164373623",
95
+ "extra_special_token_88": "676539973863",
96
+ "extra_special_token_89": "826214551218",
97
+ "extra_special_token_90": "182345464792",
98
+ "extra_special_token_91": "232776999554",
99
+ "extra_special_token_92": "337326533813",
100
+ "extra_special_token_93": "676676697292",
101
+ "extra_special_token_94": "929185622831",
102
+ "extra_special_token_95": "545512344383",
103
+ "extra_special_token_96": "499444466686",
104
+ "extra_special_token_97": "314697386682",
105
+ "extra_special_token_98": "517379856925",
106
+ "extra_special_token_99": "379557332953",
107
+ "extra_special_token_100": "614797267726",
108
+ "extra_special_token_101": "429781429464",
109
+ "extra_special_token_102": "922466849763",
110
+ "extra_special_token_103": "721737645236",
111
+ "extra_special_token_104": "479227349997",
112
+ "extra_special_token_105": "136931728327",
113
+ "extra_special_token_106": "259533577263",
114
+ "extra_special_token_107": "488538864842",
115
+ "extra_special_token_108": "937495658852",
116
+ "extra_special_token_109": "489991411364",
117
+ "extra_special_token_110": "499148455254",
118
+ "extra_special_token_111": "441373944925",
119
+ "extra_special_token_112": "899151413682",
120
+ "extra_special_token_113": "467893531755",
121
+ "extra_special_token_114": "527117488925",
122
+ "extra_special_token_115": "928335588653",
123
+ "extra_special_token_116": "374439448821",
124
+ "extra_special_token_117": "879425227932",
125
+ "extra_special_token_118": "867678158885",
126
+ "extra_special_token_119": "399749397872",
127
+ "extra_special_token_120": "129693547287",
128
+ "extra_special_token_121": "689285841825",
129
+ "extra_special_token_122": "771619544974",
130
+ "extra_special_token_123": "724883568652",
131
+ "extra_special_token_124": "516968424863",
132
+ "extra_special_token_125": "733737988257",
133
+ "extra_special_token_126": "852347289392",
134
+ "extra_special_token_127": "296953381169",
135
+ "extra_special_token_128": "377273562477",
136
+ "extra_special_token_129": "262296912232",
137
+ "extra_special_token_130": "547149832394",
138
+ "extra_special_token_131": "298464134954",
139
+ "extra_special_token_132": "216667245274",
140
+ "extra_special_token_133": "843998562287",
141
+ "extra_special_token_134": "572154333646",
142
+ "extra_special_token_135": "124589118494",
143
+ "extra_special_token_136": "841824384614",
144
+ "extra_special_token_137": "232896526252",
145
+ "extra_special_token_138": "295448593321",
146
+ "extra_special_token_139": "123741461297",
147
+ "extra_special_token_140": "653573457168",
148
+ "extra_special_token_141": "196735786156",
149
+ "extra_special_token_142": "377338713663",
150
+ "extra_special_token_143": "964342468552",
151
+ "extra_special_token_144": "586855179568",
152
+ "extra_special_token_145": "484773717614",
153
+ "extra_special_token_146": "894885246797",
154
+ "extra_special_token_147": "677896358599",
155
+ "extra_special_token_148": "848845611563",
156
+ "extra_special_token_149": "851852651677",
157
+ "extra_special_token_150": "398549545767",
158
+ "extra_special_token_151": "454244839926",
159
+ "extra_special_token_152": "799364566435",
160
+ "extra_special_token_153": "967114116556",
161
+ "extra_special_token_154": "817378986438",
162
+ "extra_special_token_155": "233795848681",
163
+ "extra_special_token_156": "824387273757",
164
+ "extra_special_token_157": "916198946615",
165
+ "extra_special_token_158": "563117729724",
166
+ "extra_special_token_159": "951794811935",
167
+ "extra_special_token_160": "374598961236",
168
+ "extra_special_token_161": "922867396683",
169
+ "extra_special_token_162": "765737843639",
170
+ "extra_special_token_163": "175469284871",
171
+ "extra_special_token_164": "231853711778",
172
+ "extra_special_token_165": "662426712668",
173
+ "extra_special_token_166": "711412347158",
174
+ "extra_special_token_167": "753466987363",
175
+ "extra_special_token_168": "513361312532",
176
+ "extra_special_token_169": "712992815957",
177
+ "extra_special_token_170": "971621888444",
178
+ "extra_special_token_171": "829235161526",
179
+ "extra_special_token_172": "585544633356",
180
+ "extra_special_token_173": "582471228164",
181
+ "extra_special_token_174": "678666359123",
182
+ "extra_special_token_175": "557533689478",
183
+ "extra_special_token_176": "632962475133",
184
+ "extra_special_token_177": "484489193824",
185
+ "extra_special_token_178": "489562189822",
186
+ "extra_special_token_179": "589547936288",
187
+ "extra_special_token_180": "363214487524",
188
+ "extra_special_token_181": "244885399387",
189
+ "extra_special_token_182": "431751228368",
190
+ "extra_special_token_183": "433581868192",
191
+ "extra_special_token_184": "486391569221",
192
+ "extra_special_token_185": "185438575221",
193
+ "extra_special_token_186": "126574388585",
194
+ "extra_special_token_187": "741757479784",
195
+ "extra_special_token_188": "529854679937",
196
+ "extra_special_token_189": "996116119839",
197
+ "extra_special_token_190": "616248973917",
198
+ "extra_special_token_191": "763531783491",
199
+ "extra_special_token_192": "955456118295",
200
+ "extra_special_token_193": "364196983365",
201
+ "extra_special_token_194": "195792996468",
202
+ "extra_special_token_195": "151859598873",
203
+ "extra_special_token_196": "399223169721",
204
+ "extra_special_token_197": "938488813964",
205
+ "extra_special_token_198": "961981959227",
206
+ "extra_special_token_199": "183368827562",
207
+ "extra_special_token_200": "533417736566",
208
+ "extra_special_token_201": "786391632558",
209
+ "extra_special_token_202": "665661658354",
210
+ "extra_special_token_203": "693281533643",
211
+ "extra_special_token_204": "475794684356",
212
+ "extra_special_token_205": "652154162978",
213
+ "extra_special_token_206": "753233719644",
214
+ "extra_special_token_207": "668514843129",
215
+ "extra_special_token_208": "819162623892",
216
+ "extra_special_token_209": "941169431859",
217
+ "extra_special_token_210": "877385381798",
218
+ "extra_special_token_211": "752644929761",
219
+ "extra_special_token_212": "881136466196",
220
+ "extra_special_token_213": "275597777299",
221
+ "extra_special_token_214": "731681792655",
222
+ "extra_special_token_215": "961133895172",
223
+ "extra_special_token_216": "864718285734",
224
+ "extra_special_token_217": "963852916563",
225
+ "extra_special_token_218": "319584985416",
226
+ "extra_special_token_219": "563365646341",
227
+ "extra_special_token_220": "811371928234",
228
+ "extra_special_token_221": "837131396371",
229
+ "extra_special_token_222": "267514771964",
230
+ "extra_special_token_223": "944513428457",
231
+ "extra_special_token_224": "117298239631",
232
+ "extra_special_token_225": "158142752582",
233
+ "extra_special_token_226": "252867443568",
234
+ "extra_special_token_227": "839269684865",
235
+ "extra_special_token_228": "612788593128",
236
+ "extra_special_token_229": "145669731981",
237
+ "extra_special_token_230": "121557291859",
238
+ "extra_special_token_231": "245416776926",
239
+ "extra_special_token_232": "799417897197",
240
+ "extra_special_token_233": "997958836435",
241
+ "extra_special_token_234": "892336777248",
242
+ "extra_special_token_235": "158929292238",
243
+ "extra_special_token_236": "581976444672",
244
+ "extra_special_token_237": "897784492783",
245
+ "extra_special_token_238": "492373714791",
246
+ "extra_special_token_239": "512659818733",
247
+ "extra_special_token_240": "881112998642",
248
+ "extra_special_token_241": "619454958782",
249
+ "extra_special_token_242": "431149748713",
250
+ "extra_special_token_243": "624221476921",
251
+ "extra_special_token_244": "125866399464",
252
+ "extra_special_token_245": "339882449689",
253
+ "extra_special_token_246": "186198784585",
254
+ "extra_special_token_247": "943193294691",
255
+ "extra_special_token_248": "955668961269",
256
+ "extra_special_token_249": "232787996724",
257
+ "extra_special_token_250": "215671314196",
258
+ "extra_special_token_251": "286173241916",
259
+ "extra_special_token_252": "745977673725",
260
+ "extra_special_token_253": "556976448182",
261
+ "extra_special_token_254": "599961512792",
262
+ "extra_special_token_255": "766294538337",
263
+ "extra_special_token_256": "934912591213",
264
+ "extra_special_token_257": "295118729589",
265
+ "extra_special_token_258": "529455466433",
266
+ "extra_special_token_259": "196119929397",
267
+ "extra_special_token_260": "379571934299",
268
+ "extra_special_token_261": "251789649997",
269
+ "extra_special_token_262": "564544131355",
270
+ "extra_special_token_263": "244371196654",
271
+ "extra_special_token_264": "384598329253",
272
+ "extra_special_token_265": "887753195844",
273
+ "extra_special_token_266": "364947325679",
274
+ "extra_special_token_267": "655517954651",
275
+ "extra_special_token_268": "673948786567",
276
+ "extra_special_token_269": "857231548835",
277
+ "extra_special_token_270": "816115936673",
278
+ "extra_special_token_271": "644234165531",
279
+ "extra_special_token_272": "182782912224",
280
+ "extra_special_token_273": "234316622259",
281
+ "extra_special_token_274": "421369185549",
282
+ "extra_special_token_275": "434632855397",
283
+ "extra_special_token_276": "921889371893",
284
+ "extra_special_token_277": "415956914763",
285
+ "extra_special_token_278": "598916996413",
286
+ "extra_special_token_279": "773671349113",
287
+ "extra_special_token_280": "952465217972",
288
+ "extra_special_token_281": "117657531962",
289
+ "extra_special_token_282": "729825168745",
290
+ "extra_special_token_283": "691315125346",
291
+ "extra_special_token_284": "768461952319",
292
+ "extra_special_token_285": "664847713559",
293
+ "extra_special_token_286": "953267689786",
294
+ "extra_special_token_287": "886464195129",
295
+ "extra_special_token_288": "824488329416",
296
+ "extra_special_token_289": "837873762491",
297
+ "extra_special_token_290": "532833541879",
298
+ "extra_special_token_291": "669183782449",
299
+ "extra_special_token_292": "941976537588",
300
+ "extra_special_token_293": "739394546916",
301
+ "extra_special_token_294": "267954879268",
302
+ "extra_special_token_295": "637551427887",
303
+ "extra_special_token_296": "217756494954",
304
+ "extra_special_token_297": "524444658383",
305
+ "extra_special_token_298": "117783274348",
306
+ "extra_special_token_299": "138218735276",
307
+ "extra_special_token_300": "814611949491",
308
+ "extra_special_token_301": "711641973413",
309
+ "extra_special_token_302": "499156317423",
310
+ "extra_special_token_303": "515856611931",
311
+ "extra_special_token_304": "454164859837",
312
+ "extra_special_token_305": "345271433112",
313
+ "extra_special_token_306": "462294118988",
314
+ "extra_special_token_307": "511785788222",
315
+ "extra_special_token_308": "497294727353",
316
+ "extra_special_token_309": "866519986723",
317
+ "extra_special_token_310": "334513529294",
318
+ "extra_special_token_311": "549946382131",
319
+ "extra_special_token_312": "284445431422",
320
+ "extra_special_token_313": "396521188476",
321
+ "extra_special_token_314": "421435255895",
322
+ "extra_special_token_315": "133373659361",
323
+ "extra_special_token_316": "322683334381",
324
+ "extra_special_token_317": "228358422847",
325
+ "extra_special_token_318": "291762694874",
326
+ "extra_special_token_319": "143182978129",
327
+ "extra_special_token_320": "511923256573",
328
+ "extra_special_token_321": "327158398268",
329
+ "extra_special_token_322": "879764613759",
330
+ "extra_special_token_323": "564395222747",
331
+ "extra_special_token_324": "451161679736",
332
+ "extra_special_token_325": "538631466654",
333
+ "extra_special_token_326": "221762325616",
334
+ "extra_special_token_327": "218391991184",
335
+ "extra_special_token_328": "322589379462",
336
+ "extra_special_token_329": "876537814263",
337
+ "extra_special_token_330": "152676556624",
338
+ "extra_special_token_331": "332522971941",
339
+ "extra_special_token_332": "884354318946",
340
+ "extra_special_token_333": "513349618943",
341
+ "extra_special_token_334": "116639746413",
342
+ "extra_special_token_335": "635185846287",
343
+ "extra_special_token_336": "993832498489",
344
+ "extra_special_token_337": "813981174797",
345
+ "extra_special_token_338": "438745114173",
346
+ "extra_special_token_339": "983493951323",
347
+ "extra_special_token_340": "724492262421",
348
+ "extra_special_token_341": "622553389126",
349
+ "extra_special_token_342": "889965243135",
350
+ "extra_special_token_343": "364492359246",
351
+ "extra_special_token_344": "154962668224",
352
+ "extra_special_token_345": "179564995814",
353
+ "extra_special_token_346": "418412875665",
354
+ "extra_special_token_347": "718951851413",
355
+ "extra_special_token_348": "699446724178",
356
+ "extra_special_token_349": "624266421831",
357
+ "extra_special_token_350": "815458725125",
358
+ "extra_special_token_351": "455423278865",
359
+ "extra_special_token_352": "393741199486",
360
+ "extra_special_token_353": "328552864359",
361
+ "extra_special_token_354": "211662639865",
362
+ "extra_special_token_355": "218784516525",
363
+ "extra_special_token_356": "762486672996",
364
+ "extra_special_token_357": "142799718159",
365
+ "extra_special_token_358": "858146415154",
366
+ "extra_special_token_359": "767858144912",
367
+ "extra_special_token_360": "571317457151",
368
+ "extra_special_token_361": "635127952696",
369
+ "extra_special_token_362": "116427191984",
370
+ "extra_special_token_363": "268921994538",
371
+ "extra_special_token_364": "523937669294",
372
+ "extra_special_token_365": "165429152138",
373
+ "extra_special_token_366": "739246183345",
374
+ "extra_special_token_367": "591464355756",
375
+ "extra_special_token_368": "212985874612",
376
+ "extra_special_token_369": "191887635211",
377
+ "extra_special_token_370": "967214577653",
378
+ "extra_special_token_371": "119342152414",
379
+ "extra_special_token_372": "946444632795",
380
+ "extra_special_token_373": "618423867817",
381
+ "extra_special_token_374": "228565148417",
382
+ "extra_special_token_375": "729116422489",
383
+ "extra_special_token_376": "527874729936",
384
+ "extra_special_token_377": "739784153482",
385
+ "extra_special_token_378": "387763951128",
386
+ "extra_special_token_379": "331369926711",
387
+ "extra_special_token_380": "562716493614",
388
+ "extra_special_token_381": "739667844957",
389
+ "extra_special_token_382": "562389434565",
390
+ "extra_special_token_383": "256497188281",
391
+ "extra_special_token_384": "859927364588",
392
+ "extra_special_token_385": "417668946583",
393
+ "extra_special_token_386": "357621613582",
394
+ "extra_special_token_387": "438435178228",
395
+ "extra_special_token_388": "485692541169",
396
+ "extra_special_token_389": "825815739116",
397
+ "extra_special_token_390": "342221452223",
398
+ "extra_special_token_391": "697747991249",
399
+ "extra_special_token_392": "716763689965",
400
+ "extra_special_token_393": "141499982867",
401
+ "extra_special_token_394": "818479319499",
402
+ "extra_special_token_395": "336813343298",
403
+ "extra_special_token_396": "594688742928",
404
+ "extra_special_token_397": "472129283475",
405
+ "extra_special_token_398": "514354144759",
406
+ "extra_special_token_399": "349249721685",
407
+ "extra_special_token_400": "546276298359",
408
+ "extra_special_token_401": "353755529131",
409
+ "extra_special_token_402": "315534574435",
410
+ "extra_special_token_403": "523723475786",
411
+ "extra_special_token_404": "215826764872",
412
+ "extra_special_token_405": "367968398551",
413
+ "extra_special_token_406": "569853653352",
414
+ "extra_special_token_407": "389715484387",
415
+ "extra_special_token_408": "293847485454",
416
+ "extra_special_token_409": "714738141818",
417
+ "extra_special_token_410": "178478368922",
418
+ "extra_special_token_411": "581493616981",
419
+ "extra_special_token_412": "589439538674",
420
+ "extra_special_token_413": "846657726193",
421
+ "extra_special_token_414": "722339992679",
422
+ "extra_special_token_415": "138154781148",
423
+ "extra_special_token_416": "757785319772",
424
+ "extra_special_token_417": "492516914298",
425
+ "extra_special_token_418": "919181521716",
426
+ "extra_special_token_419": "985781138935",
427
+ "extra_special_token_420": "476969195485",
428
+ "extra_special_token_421": "313145133463",
429
+ "extra_special_token_422": "758963111966",
430
+ "extra_special_token_423": "147541537162",
431
+ "extra_special_token_424": "557163366873",
432
+ "extra_special_token_425": "144373897488",
433
+ "extra_special_token_426": "522515164754",
434
+ "extra_special_token_427": "724964923582",
435
+ "extra_special_token_428": "284776712475",
436
+ "extra_special_token_429": "375429755114",
437
+ "extra_special_token_430": "181233596124",
438
+ "extra_special_token_431": "948585673431",
439
+ "extra_special_token_432": "243165586174",
440
+ "extra_special_token_433": "396847976144",
441
+ "extra_special_token_434": "997724962668",
442
+ "extra_special_token_435": "558837194455",
443
+ "extra_special_token_436": "163165456396",
444
+ "extra_special_token_437": "378749551722",
445
+ "extra_special_token_438": "161238482259",
446
+ "extra_special_token_439": "754978243758",
447
+ "extra_special_token_440": "195388849133",
448
+ "extra_special_token_441": "229775525672",
449
+ "extra_special_token_442": "262437452884",
450
+ "extra_special_token_443": "441377892146",
451
+ "extra_special_token_444": "451885565366",
452
+ "extra_special_token_445": "981277526855",
453
+ "extra_special_token_446": "762495822823",
454
+ "extra_special_token_447": "368763327262",
455
+ "extra_special_token_448": "757422791351",
456
+ "extra_special_token_449": "636324136426",
457
+ "extra_special_token_450": "214193645583",
458
+ "extra_special_token_451": "412843856172",
459
+ "extra_special_token_452": "179386156569",
460
+ "extra_special_token_453": "756916173536",
461
+ "extra_special_token_454": "892697125149",
462
+ "extra_special_token_455": "625334487352",
463
+ "extra_special_token_456": "941861857715",
464
+ "extra_special_token_457": "887417525236",
465
+ "extra_special_token_458": "649516938598",
466
+ "extra_special_token_459": "717628619782",
467
+ "extra_special_token_460": "438124184139",
468
+ "extra_special_token_461": "547563892268",
469
+ "extra_special_token_462": "856317483891",
470
+ "extra_special_token_463": "313313831273",
471
+ "extra_special_token_464": "371496153876",
472
+ "extra_special_token_465": "587541149322",
473
+ "extra_special_token_466": "265847332563",
474
+ "extra_special_token_467": "449549215429",
475
+ "extra_special_token_468": "163497196769",
476
+ "extra_special_token_469": "861342291298",
477
+ "extra_special_token_470": "268433315926",
478
+ "extra_special_token_471": "774679513717",
479
+ "extra_special_token_472": "851254219729",
480
+ "extra_special_token_473": "583527834464",
481
+ "extra_special_token_474": "488496781997",
482
+ "extra_special_token_475": "556814553861",
483
+ "extra_special_token_476": "482829231639",
484
+ "extra_special_token_477": "618878266619",
485
+ "extra_special_token_478": "147444452794",
486
+ "extra_special_token_479": "949235426629",
487
+ "extra_special_token_480": "357299947518",
488
+ "extra_special_token_481": "175528632226",
489
+ "extra_special_token_482": "645527857972",
490
+ "extra_special_token_483": "186872457894",
491
+ "extra_special_token_484": "552738847828",
492
+ "extra_special_token_485": "626748382482",
493
+ "extra_special_token_486": "921894985642",
494
+ "extra_special_token_487": "943878645871",
495
+ "extra_special_token_488": "859289776479",
496
+ "extra_special_token_489": "614583493135",
497
+ "extra_special_token_490": "933775286797",
498
+ "extra_special_token_491": "332234613346",
499
+ "extra_special_token_492": "325196781219",
500
+ "extra_special_token_493": "142526557681",
501
+ "extra_special_token_494": "356722692178",
502
+ "extra_special_token_495": "449318681694",
503
+ "extra_special_token_496": "687284547244",
504
+ "extra_special_token_497": "947262995132",
505
+ "extra_special_token_498": "893974619684",
506
+ "extra_special_token_499": "797238311233"
507
+ },
508
+ "is_local": false,
509
+ "model_input_names": [
510
+ "input_ids",
511
+ "attention_mask"
512
+ ],
513
+ "model_max_length": 131072,
514
+ "pad_token": "<|end_of_text|>",
515
+ "tokenizer_class": "PreTrainedTokenizerFast"
516
+ }
llama3.1-8b/dp8/train.log ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-03-29 08:18:38,126 [INFO] new_opacus_codex.train_steps: epoch=1 step=5 loss=12.7582
2
+ 2026-03-29 08:19:21,570 [INFO] new_opacus_codex.train_steps: epoch=1 step=10 loss=12.7552
3
+ 2026-03-29 08:19:35,828 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=10 eval_loss=13.0680 duration_sec=13.61
4
+ 2026-03-29 08:20:18,976 [INFO] new_opacus_codex.train_steps: epoch=1 step=15 loss=12.7467
5
+ 2026-03-29 08:21:02,159 [INFO] new_opacus_codex.train_steps: epoch=1 step=20 loss=12.6976
6
+ 2026-03-29 08:21:15,985 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=20 eval_loss=13.0337 duration_sec=13.82
7
+ 2026-03-29 08:21:58,691 [INFO] new_opacus_codex.train_steps: epoch=1 step=25 loss=12.7483
8
+ 2026-03-29 08:22:41,007 [INFO] new_opacus_codex.train_steps: epoch=1 step=30 loss=12.7280
9
+ 2026-03-29 08:22:54,833 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=30 eval_loss=12.9749 duration_sec=13.81
10
+ 2026-03-29 08:23:37,932 [INFO] new_opacus_codex.train_steps: epoch=1 step=35 loss=12.6013
11
+ 2026-03-29 08:24:20,561 [INFO] new_opacus_codex.train_steps: epoch=1 step=40 loss=12.7136
12
+ 2026-03-29 08:24:34,395 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=40 eval_loss=12.9016 duration_sec=13.82
13
+ 2026-03-29 08:25:17,014 [INFO] new_opacus_codex.train_steps: epoch=1 step=45 loss=12.7233
14
+ 2026-03-29 08:26:00,492 [INFO] new_opacus_codex.train_steps: epoch=1 step=50 loss=12.5865
15
+ 2026-03-29 08:26:14,575 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=50 eval_loss=12.8164 duration_sec=13.47
16
+ 2026-03-29 08:26:57,223 [INFO] new_opacus_codex.train_steps: epoch=1 step=55 loss=12.5746
17
+ 2026-03-29 08:27:39,547 [INFO] new_opacus_codex.train_steps: epoch=1 step=60 loss=12.5392
18
+ 2026-03-29 08:27:53,386 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=60 eval_loss=12.7162 duration_sec=13.76
19
+ 2026-03-29 08:28:35,518 [INFO] new_opacus_codex.train_steps: epoch=1 step=65 loss=12.5447
20
+ 2026-03-29 08:29:18,831 [INFO] new_opacus_codex.train_steps: epoch=1 step=70 loss=12.5568
21
+ 2026-03-29 08:29:32,674 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=70 eval_loss=12.6048 duration_sec=13.76
22
+ 2026-03-29 08:30:16,231 [INFO] new_opacus_codex.train_steps: epoch=1 step=75 loss=12.5089
23
+ 2026-03-29 08:30:59,618 [INFO] new_opacus_codex.train_steps: epoch=1 step=80 loss=12.3639
24
+ 2026-03-29 08:31:13,575 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=80 eval_loss=12.4824 duration_sec=13.56
25
+ 2026-03-29 08:31:57,708 [INFO] new_opacus_codex.train_steps: epoch=1 step=85 loss=12.2419
26
+ 2026-03-29 08:32:41,058 [INFO] new_opacus_codex.train_steps: epoch=1 step=90 loss=12.1949
27
+ 2026-03-29 08:32:54,899 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=90 eval_loss=12.3487 duration_sec=13.80
28
+ 2026-03-29 08:33:37,588 [INFO] new_opacus_codex.train_steps: epoch=1 step=95 loss=12.2511
29
+ 2026-03-29 08:34:20,398 [INFO] new_opacus_codex.train_steps: epoch=1 step=100 loss=12.2906
30
+ 2026-03-29 08:34:34,264 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=100 eval_loss=12.1817 duration_sec=13.73
31
+ 2026-03-29 08:35:18,849 [INFO] new_opacus_codex.train_steps: epoch=1 step=105 loss=12.1571
32
+ 2026-03-29 08:36:01,113 [INFO] new_opacus_codex.train_steps: epoch=1 step=110 loss=12.1120
33
+ 2026-03-29 08:36:14,939 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=110 eval_loss=11.9971 duration_sec=13.67
34
+ 2026-03-29 08:36:58,037 [INFO] new_opacus_codex.train_steps: epoch=1 step=115 loss=11.9682
35
+ 2026-03-29 08:37:40,230 [INFO] new_opacus_codex.train_steps: epoch=1 step=120 loss=11.8005
36
+ 2026-03-29 08:37:54,071 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=120 eval_loss=11.7866 duration_sec=13.83
37
+ 2026-03-29 08:38:36,314 [INFO] new_opacus_codex.train_steps: epoch=1 step=125 loss=11.8008
38
+ 2026-03-29 08:39:20,204 [INFO] new_opacus_codex.train_steps: epoch=1 step=130 loss=11.6446
39
+ 2026-03-29 08:39:34,066 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=130 eval_loss=11.5608 duration_sec=13.85
40
+ 2026-03-29 08:40:17,140 [INFO] new_opacus_codex.train_steps: epoch=1 step=135 loss=11.5087
41
+ 2026-03-29 08:40:58,984 [INFO] new_opacus_codex.train_steps: epoch=1 step=140 loss=11.5181
42
+ 2026-03-29 08:41:12,816 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=140 eval_loss=11.3173 duration_sec=13.63
43
+ 2026-03-29 08:41:55,575 [INFO] new_opacus_codex.train_steps: epoch=1 step=145 loss=11.3733
44
+ 2026-03-29 08:42:39,197 [INFO] new_opacus_codex.train_steps: epoch=1 step=150 loss=11.2248
45
+ 2026-03-29 08:42:53,038 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=150 eval_loss=11.0668 duration_sec=13.79
46
+ 2026-03-29 08:43:35,478 [INFO] new_opacus_codex.train_steps: epoch=1 step=155 loss=11.0097
47
+ 2026-03-29 08:44:16,836 [INFO] new_opacus_codex.train_steps: epoch=1 step=160 loss=10.9057
48
+ 2026-03-29 08:44:30,688 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=160 eval_loss=10.8182 duration_sec=13.84
49
+ 2026-03-29 08:45:14,025 [INFO] new_opacus_codex.train_steps: epoch=1 step=165 loss=10.9365
50
+ 2026-03-29 08:45:57,664 [INFO] new_opacus_codex.train_steps: epoch=1 step=170 loss=10.7913
51
+ 2026-03-29 08:46:11,533 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=170 eval_loss=10.5750 duration_sec=13.86
52
+ 2026-03-29 08:46:54,127 [INFO] new_opacus_codex.train_steps: epoch=1 step=175 loss=10.6257
53
+ 2026-03-29 08:47:37,956 [INFO] new_opacus_codex.train_steps: epoch=1 step=180 loss=10.6356
54
+ 2026-03-29 08:47:51,787 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=1 step=180 eval_loss=10.3357 duration_sec=13.76
55
+ 2026-03-29 08:48:57,093 [INFO] new_opacus_codex.train_steps: epoch=2 step=185 loss=10.7682
56
+ 2026-03-29 08:49:42,151 [INFO] new_opacus_codex.train_steps: epoch=2 step=190 loss=10.4391
57
+ 2026-03-29 08:49:55,971 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=190 eval_loss=10.1074 duration_sec=13.81
58
+ 2026-03-29 08:50:39,635 [INFO] new_opacus_codex.train_steps: epoch=2 step=195 loss=10.3149
59
+ 2026-03-29 08:51:22,864 [INFO] new_opacus_codex.train_steps: epoch=2 step=200 loss=10.1479
60
+ 2026-03-29 08:51:36,706 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=200 eval_loss=9.8868 duration_sec=13.83
61
+ 2026-03-29 08:52:19,996 [INFO] new_opacus_codex.train_steps: epoch=2 step=205 loss=10.0115
62
+ 2026-03-29 08:53:03,261 [INFO] new_opacus_codex.train_steps: epoch=2 step=210 loss=9.9104
63
+ 2026-03-29 08:53:17,119 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=210 eval_loss=9.6749 duration_sec=13.85
64
+ 2026-03-29 08:53:59,720 [INFO] new_opacus_codex.train_steps: epoch=2 step=215 loss=9.8469
65
+ 2026-03-29 08:54:42,729 [INFO] new_opacus_codex.train_steps: epoch=2 step=220 loss=9.8212
66
+ 2026-03-29 08:54:56,651 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=220 eval_loss=9.4871 duration_sec=13.91
67
+ 2026-03-29 08:55:40,194 [INFO] new_opacus_codex.train_steps: epoch=2 step=225 loss=9.6758
68
+ 2026-03-29 08:56:22,824 [INFO] new_opacus_codex.train_steps: epoch=2 step=230 loss=9.6581
69
+ 2026-03-29 08:56:36,708 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=230 eval_loss=9.3206 duration_sec=13.86
70
+ 2026-03-29 08:57:18,725 [INFO] new_opacus_codex.train_steps: epoch=2 step=235 loss=9.7477
71
+ 2026-03-29 08:58:02,240 [INFO] new_opacus_codex.train_steps: epoch=2 step=240 loss=9.6920
72
+ 2026-03-29 08:58:16,076 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=240 eval_loss=9.1688 duration_sec=13.80
73
+ 2026-03-29 08:58:59,224 [INFO] new_opacus_codex.train_steps: epoch=2 step=245 loss=9.4260
74
+ 2026-03-29 08:59:42,581 [INFO] new_opacus_codex.train_steps: epoch=2 step=250 loss=9.3598
75
+ 2026-03-29 08:59:56,450 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=250 eval_loss=9.0314 duration_sec=13.85
76
+ 2026-03-29 09:00:39,700 [INFO] new_opacus_codex.train_steps: epoch=2 step=255 loss=9.3229
77
+ 2026-03-29 09:01:22,066 [INFO] new_opacus_codex.train_steps: epoch=2 step=260 loss=9.1423
78
+ 2026-03-29 09:01:35,940 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=260 eval_loss=8.9018 duration_sec=13.82
79
+ 2026-03-29 09:02:19,031 [INFO] new_opacus_codex.train_steps: epoch=2 step=265 loss=9.1008
80
+ 2026-03-29 09:03:01,072 [INFO] new_opacus_codex.train_steps: epoch=2 step=270 loss=9.1319
81
+ 2026-03-29 09:03:14,939 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=270 eval_loss=8.7856 duration_sec=13.85
82
+ 2026-03-29 09:03:58,377 [INFO] new_opacus_codex.train_steps: epoch=2 step=275 loss=9.1451
83
+ 2026-03-29 09:04:43,141 [INFO] new_opacus_codex.train_steps: epoch=2 step=280 loss=9.0250
84
+ 2026-03-29 09:04:56,992 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=280 eval_loss=8.6834 duration_sec=13.70
85
+ 2026-03-29 09:05:39,648 [INFO] new_opacus_codex.train_steps: epoch=2 step=285 loss=8.9475
86
+ 2026-03-29 09:06:23,051 [INFO] new_opacus_codex.train_steps: epoch=2 step=290 loss=8.9228
87
+ 2026-03-29 09:06:36,918 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=290 eval_loss=8.5970 duration_sec=13.83
88
+ 2026-03-29 09:07:20,686 [INFO] new_opacus_codex.train_steps: epoch=2 step=295 loss=8.8888
89
+ 2026-03-29 09:08:03,321 [INFO] new_opacus_codex.train_steps: epoch=2 step=300 loss=8.9310
90
+ 2026-03-29 09:08:17,163 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=300 eval_loss=8.5208 duration_sec=13.83
91
+ 2026-03-29 09:09:01,040 [INFO] new_opacus_codex.train_steps: epoch=2 step=305 loss=8.9299
92
+ 2026-03-29 09:09:44,340 [INFO] new_opacus_codex.train_steps: epoch=2 step=310 loss=8.8965
93
+ 2026-03-29 09:09:58,334 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=310 eval_loss=8.4586 duration_sec=13.85
94
+ 2026-03-29 09:10:42,263 [INFO] new_opacus_codex.train_steps: epoch=2 step=315 loss=8.8381
95
+ 2026-03-29 09:11:26,111 [INFO] new_opacus_codex.train_steps: epoch=2 step=320 loss=8.8937
96
+ 2026-03-29 09:11:40,011 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=320 eval_loss=8.4036 duration_sec=13.84
97
+ 2026-03-29 09:12:23,152 [INFO] new_opacus_codex.train_steps: epoch=2 step=325 loss=8.8625
98
+ 2026-03-29 09:13:05,819 [INFO] new_opacus_codex.train_steps: epoch=2 step=330 loss=8.8267
99
+ 2026-03-29 09:13:19,650 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=330 eval_loss=8.3551 duration_sec=13.73
100
+ 2026-03-29 09:14:01,448 [INFO] new_opacus_codex.train_steps: epoch=2 step=335 loss=8.7488
101
+ 2026-03-29 09:14:44,754 [INFO] new_opacus_codex.train_steps: epoch=2 step=340 loss=8.5471
102
+ 2026-03-29 09:14:58,632 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=340 eval_loss=8.3166 duration_sec=13.86
103
+ 2026-03-29 09:15:41,109 [INFO] new_opacus_codex.train_steps: epoch=2 step=345 loss=8.4518
104
+ 2026-03-29 09:16:24,890 [INFO] new_opacus_codex.train_steps: epoch=2 step=350 loss=8.4608
105
+ 2026-03-29 09:16:38,829 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=350 eval_loss=8.2862 duration_sec=13.90
106
+ 2026-03-29 09:17:21,170 [INFO] new_opacus_codex.train_steps: epoch=2 step=355 loss=8.6859
107
+ 2026-03-29 09:18:03,767 [INFO] new_opacus_codex.train_steps: epoch=2 step=360 loss=8.8237
108
+ 2026-03-29 09:18:17,614 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=2 step=360 eval_loss=8.2570 duration_sec=13.82
109
+ 2026-03-29 09:18:59,880 [INFO] new_opacus_codex.train_steps: epoch=2 step=365 loss=8.7437
110
+ 2026-03-29 09:20:06,380 [INFO] new_opacus_codex.train_steps: epoch=3 step=370 loss=8.6091
111
+ 2026-03-29 09:20:20,203 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=370 eval_loss=8.2279 duration_sec=13.80
112
+ 2026-03-29 09:21:03,844 [INFO] new_opacus_codex.train_steps: epoch=3 step=375 loss=8.3830
113
+ 2026-03-29 09:21:47,792 [INFO] new_opacus_codex.train_steps: epoch=3 step=380 loss=8.4251
114
+ 2026-03-29 09:22:01,631 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=380 eval_loss=8.2055 duration_sec=13.75
115
+ 2026-03-29 09:22:43,921 [INFO] new_opacus_codex.train_steps: epoch=3 step=385 loss=8.5817
116
+ 2026-03-29 09:23:27,219 [INFO] new_opacus_codex.train_steps: epoch=3 step=390 loss=8.5799
117
+ 2026-03-29 09:23:41,063 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=390 eval_loss=8.1823 duration_sec=13.81
118
+ 2026-03-29 09:24:23,682 [INFO] new_opacus_codex.train_steps: epoch=3 step=395 loss=8.5398
119
+ 2026-03-29 09:25:07,472 [INFO] new_opacus_codex.train_steps: epoch=3 step=400 loss=8.4981
120
+ 2026-03-29 09:25:21,293 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=400 eval_loss=8.1601 duration_sec=13.78
121
+ 2026-03-29 09:26:03,805 [INFO] new_opacus_codex.train_steps: epoch=3 step=405 loss=8.4369
122
+ 2026-03-29 09:26:47,933 [INFO] new_opacus_codex.train_steps: epoch=3 step=410 loss=8.5627
123
+ 2026-03-29 09:27:01,770 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=410 eval_loss=8.1336 duration_sec=13.76
124
+ 2026-03-29 09:27:44,244 [INFO] new_opacus_codex.train_steps: epoch=3 step=415 loss=8.5231
125
+ 2026-03-29 09:28:28,187 [INFO] new_opacus_codex.train_steps: epoch=3 step=420 loss=8.4865
126
+ 2026-03-29 09:28:42,101 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=420 eval_loss=8.1086 duration_sec=13.83
127
+ 2026-03-29 09:29:26,108 [INFO] new_opacus_codex.train_steps: epoch=3 step=425 loss=8.4527
128
+ 2026-03-29 09:30:09,157 [INFO] new_opacus_codex.train_steps: epoch=3 step=430 loss=8.3991
129
+ 2026-03-29 09:30:23,057 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=430 eval_loss=8.0876 duration_sec=13.82
130
+ 2026-03-29 09:31:06,280 [INFO] new_opacus_codex.train_steps: epoch=3 step=435 loss=8.3516
131
+ 2026-03-29 09:31:48,249 [INFO] new_opacus_codex.train_steps: epoch=3 step=440 loss=8.2078
132
+ 2026-03-29 09:32:02,137 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=440 eval_loss=8.0669 duration_sec=13.86
133
+ 2026-03-29 09:32:46,192 [INFO] new_opacus_codex.train_steps: epoch=3 step=445 loss=8.3159
134
+ 2026-03-29 09:33:28,921 [INFO] new_opacus_codex.train_steps: epoch=3 step=450 loss=8.4696
135
+ 2026-03-29 09:33:42,729 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=450 eval_loss=8.0462 duration_sec=13.64
136
+ 2026-03-29 09:34:25,921 [INFO] new_opacus_codex.train_steps: epoch=3 step=455 loss=8.5669
137
+ 2026-03-29 09:35:08,904 [INFO] new_opacus_codex.train_steps: epoch=3 step=460 loss=8.4787
138
+ 2026-03-29 09:35:22,799 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=460 eval_loss=8.0277 duration_sec=13.73
139
+ 2026-03-29 09:36:05,442 [INFO] new_opacus_codex.train_steps: epoch=3 step=465 loss=8.4585
140
+ 2026-03-29 09:36:48,262 [INFO] new_opacus_codex.train_steps: epoch=3 step=470 loss=8.4779
141
+ 2026-03-29 09:37:02,110 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=470 eval_loss=8.0104 duration_sec=13.78
142
+ 2026-03-29 09:37:45,166 [INFO] new_opacus_codex.train_steps: epoch=3 step=475 loss=8.4868
143
+ 2026-03-29 09:38:27,805 [INFO] new_opacus_codex.train_steps: epoch=3 step=480 loss=8.4952
144
+ 2026-03-29 09:38:41,683 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=480 eval_loss=7.9926 duration_sec=13.84
145
+ 2026-03-29 09:39:24,404 [INFO] new_opacus_codex.train_steps: epoch=3 step=485 loss=8.3578
146
+ 2026-03-29 09:40:08,288 [INFO] new_opacus_codex.train_steps: epoch=3 step=490 loss=8.4075
147
+ 2026-03-29 09:40:22,171 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=490 eval_loss=7.9754 duration_sec=13.78
148
+ 2026-03-29 09:41:04,949 [INFO] new_opacus_codex.train_steps: epoch=3 step=495 loss=8.5507
149
+ 2026-03-29 09:41:48,750 [INFO] new_opacus_codex.train_steps: epoch=3 step=500 loss=8.5192
150
+ 2026-03-29 09:42:02,683 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=500 eval_loss=7.9597 duration_sec=13.68
151
+ 2026-03-29 09:42:45,673 [INFO] new_opacus_codex.train_steps: epoch=3 step=505 loss=8.3974
152
+ 2026-03-29 09:43:29,124 [INFO] new_opacus_codex.train_steps: epoch=3 step=510 loss=8.4250
153
+ 2026-03-29 09:43:42,954 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=510 eval_loss=7.9442 duration_sec=13.79
154
+ 2026-03-29 09:44:25,921 [INFO] new_opacus_codex.train_steps: epoch=3 step=515 loss=8.3495
155
+ 2026-03-29 09:45:09,257 [INFO] new_opacus_codex.train_steps: epoch=3 step=520 loss=8.4002
156
+ 2026-03-29 09:45:23,155 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=520 eval_loss=7.9299 duration_sec=13.80
157
+ 2026-03-29 09:46:04,810 [INFO] new_opacus_codex.train_steps: epoch=3 step=525 loss=8.3650
158
+ 2026-03-29 09:46:48,188 [INFO] new_opacus_codex.train_steps: epoch=3 step=530 loss=8.1285
159
+ 2026-03-29 09:47:02,021 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=530 eval_loss=7.9160 duration_sec=13.71
160
+ 2026-03-29 09:47:45,859 [INFO] new_opacus_codex.train_steps: epoch=3 step=535 loss=8.3950
161
+ 2026-03-29 09:48:29,555 [INFO] new_opacus_codex.train_steps: epoch=3 step=540 loss=8.3824
162
+ 2026-03-29 09:48:43,454 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=540 eval_loss=7.9028 duration_sec=13.83
163
+ 2026-03-29 09:49:26,541 [INFO] new_opacus_codex.train_steps: epoch=3 step=545 loss=8.2507
164
+ 2026-03-29 09:50:09,694 [INFO] new_opacus_codex.train_steps: epoch=3 step=550 loss=8.2800
165
+ 2026-03-29 09:50:23,505 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=3 step=550 eval_loss=7.8892 duration_sec=13.73
166
+ 2026-03-29 09:51:30,564 [INFO] new_opacus_codex.train_steps: epoch=4 step=555 loss=8.1656
167
+ 2026-03-29 09:52:15,158 [INFO] new_opacus_codex.train_steps: epoch=4 step=560 loss=8.1273
168
+ 2026-03-29 09:52:28,985 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=560 eval_loss=7.8759 duration_sec=13.75
169
+ 2026-03-29 09:53:12,139 [INFO] new_opacus_codex.train_steps: epoch=4 step=565 loss=8.1876
170
+ 2026-03-29 09:53:55,633 [INFO] new_opacus_codex.train_steps: epoch=4 step=570 loss=8.2553
171
+ 2026-03-29 09:54:09,487 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=570 eval_loss=7.8630 duration_sec=13.82
172
+ 2026-03-29 09:54:52,896 [INFO] new_opacus_codex.train_steps: epoch=4 step=575 loss=8.1838
173
+ 2026-03-29 09:55:36,594 [INFO] new_opacus_codex.train_steps: epoch=4 step=580 loss=8.1373
174
+ 2026-03-29 09:55:50,498 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=580 eval_loss=7.8497 duration_sec=13.77
175
+ 2026-03-29 09:56:34,518 [INFO] new_opacus_codex.train_steps: epoch=4 step=585 loss=8.2595
176
+ 2026-03-29 09:57:17,891 [INFO] new_opacus_codex.train_steps: epoch=4 step=590 loss=8.3531
177
+ 2026-03-29 09:57:31,782 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=590 eval_loss=7.8381 duration_sec=13.73
178
+ 2026-03-29 09:58:14,494 [INFO] new_opacus_codex.train_steps: epoch=4 step=595 loss=8.2278
179
+ 2026-03-29 09:58:59,255 [INFO] new_opacus_codex.train_steps: epoch=4 step=600 loss=8.0766
180
+ 2026-03-29 09:59:13,223 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=600 eval_loss=7.8275 duration_sec=13.83
181
+ 2026-03-29 09:59:55,943 [INFO] new_opacus_codex.train_steps: epoch=4 step=605 loss=8.2059
182
+ 2026-03-29 10:00:38,580 [INFO] new_opacus_codex.train_steps: epoch=4 step=610 loss=8.2801
183
+ 2026-03-29 10:00:52,480 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=610 eval_loss=7.8160 duration_sec=13.88
184
+ 2026-03-29 10:01:36,865 [INFO] new_opacus_codex.train_steps: epoch=4 step=615 loss=8.2740
185
+ 2026-03-29 10:02:19,581 [INFO] new_opacus_codex.train_steps: epoch=4 step=620 loss=8.3677
186
+ 2026-03-29 10:02:33,445 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=620 eval_loss=7.8050 duration_sec=13.85
187
+ 2026-03-29 10:03:16,905 [INFO] new_opacus_codex.train_steps: epoch=4 step=625 loss=8.2854
188
+ 2026-03-29 10:04:00,756 [INFO] new_opacus_codex.train_steps: epoch=4 step=630 loss=8.3025
189
+ 2026-03-29 10:04:14,630 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=630 eval_loss=7.7945 duration_sec=13.81
190
+ 2026-03-29 10:04:58,299 [INFO] new_opacus_codex.train_steps: epoch=4 step=635 loss=8.3714
191
+ 2026-03-29 10:05:42,162 [INFO] new_opacus_codex.train_steps: epoch=4 step=640 loss=8.2812
192
+ 2026-03-29 10:05:56,019 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=640 eval_loss=7.7857 duration_sec=13.83
193
+ 2026-03-29 10:06:39,488 [INFO] new_opacus_codex.train_steps: epoch=4 step=645 loss=8.2306
194
+ 2026-03-29 10:07:23,002 [INFO] new_opacus_codex.train_steps: epoch=4 step=650 loss=8.2134
195
+ 2026-03-29 10:07:36,833 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=650 eval_loss=7.7771 duration_sec=13.72
196
+ 2026-03-29 10:08:18,972 [INFO] new_opacus_codex.train_steps: epoch=4 step=655 loss=8.2380
197
+ 2026-03-29 10:09:02,512 [INFO] new_opacus_codex.train_steps: epoch=4 step=660 loss=8.0889
198
+ 2026-03-29 10:09:16,376 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=660 eval_loss=7.7696 duration_sec=13.83
199
+ 2026-03-29 10:09:59,986 [INFO] new_opacus_codex.train_steps: epoch=4 step=665 loss=8.0956
200
+ 2026-03-29 10:10:42,336 [INFO] new_opacus_codex.train_steps: epoch=4 step=670 loss=8.2664
201
+ 2026-03-29 10:10:56,222 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=670 eval_loss=7.7630 duration_sec=13.76
202
+ 2026-03-29 10:11:38,587 [INFO] new_opacus_codex.train_steps: epoch=4 step=675 loss=8.2526
203
+ 2026-03-29 10:12:21,008 [INFO] new_opacus_codex.train_steps: epoch=4 step=680 loss=8.1861
204
+ 2026-03-29 10:12:34,888 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=680 eval_loss=7.7552 duration_sec=13.68
205
+ 2026-03-29 10:13:17,669 [INFO] new_opacus_codex.train_steps: epoch=4 step=685 loss=8.0646
206
+ 2026-03-29 10:14:01,727 [INFO] new_opacus_codex.train_steps: epoch=4 step=690 loss=8.1045
207
+ 2026-03-29 10:14:15,560 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=690 eval_loss=7.7468 duration_sec=13.82
208
+ 2026-03-29 10:14:58,791 [INFO] new_opacus_codex.train_steps: epoch=4 step=695 loss=8.2821
209
+ 2026-03-29 10:15:41,821 [INFO] new_opacus_codex.train_steps: epoch=4 step=700 loss=8.2692
210
+ 2026-03-29 10:15:55,699 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=700 eval_loss=7.7379 duration_sec=13.82
211
+ 2026-03-29 10:16:37,872 [INFO] new_opacus_codex.train_steps: epoch=4 step=705 loss=8.1275
212
+ 2026-03-29 10:17:21,294 [INFO] new_opacus_codex.train_steps: epoch=4 step=710 loss=8.1881
213
+ 2026-03-29 10:17:35,204 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=710 eval_loss=7.7294 duration_sec=13.70
214
+ 2026-03-29 10:18:17,298 [INFO] new_opacus_codex.train_steps: epoch=4 step=715 loss=8.1302
215
+ 2026-03-29 10:19:00,871 [INFO] new_opacus_codex.train_steps: epoch=4 step=720 loss=8.1464
216
+ 2026-03-29 10:19:14,768 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=720 eval_loss=7.7224 duration_sec=13.86
217
+ 2026-03-29 10:19:59,775 [INFO] new_opacus_codex.train_steps: epoch=4 step=725 loss=8.1776
218
+ 2026-03-29 10:20:43,748 [INFO] new_opacus_codex.train_steps: epoch=4 step=730 loss=8.1178
219
+ 2026-03-29 10:20:57,816 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=4 step=730 eval_loss=7.7165 duration_sec=13.95
220
+ 2026-03-29 10:21:42,110 [INFO] new_opacus_codex.train_steps: epoch=4 step=735 loss=8.1309
221
+ 2026-03-29 10:22:49,209 [INFO] new_opacus_codex.train_steps: epoch=5 step=740 loss=8.2755
222
+ 2026-03-29 10:23:03,302 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=740 eval_loss=7.7103 duration_sec=14.03
223
+ 2026-03-29 10:23:47,123 [INFO] new_opacus_codex.train_steps: epoch=5 step=745 loss=8.3432
224
+ 2026-03-29 10:24:30,663 [INFO] new_opacus_codex.train_steps: epoch=5 step=750 loss=8.1016
225
+ 2026-03-29 10:24:44,786 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=750 eval_loss=7.7042 duration_sec=14.08
226
+ 2026-03-29 10:25:29,248 [INFO] new_opacus_codex.train_steps: epoch=5 step=755 loss=7.9562
227
+ 2026-03-29 10:26:11,761 [INFO] new_opacus_codex.train_steps: epoch=5 step=760 loss=8.1854
228
+ 2026-03-29 10:26:25,788 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=760 eval_loss=7.6981 duration_sec=14.01
229
+ 2026-03-29 10:27:10,147 [INFO] new_opacus_codex.train_steps: epoch=5 step=765 loss=8.1615
230
+ 2026-03-29 10:27:52,867 [INFO] new_opacus_codex.train_steps: epoch=5 step=770 loss=8.0823
231
+ 2026-03-29 10:28:06,874 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=770 eval_loss=7.6926 duration_sec=13.60
232
+ 2026-03-29 10:28:49,712 [INFO] new_opacus_codex.train_steps: epoch=5 step=775 loss=8.1717
233
+ 2026-03-29 10:29:33,629 [INFO] new_opacus_codex.train_steps: epoch=5 step=780 loss=8.3214
234
+ 2026-03-29 10:29:47,667 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=780 eval_loss=7.6876 duration_sec=14.00
235
+ 2026-03-29 10:30:31,151 [INFO] new_opacus_codex.train_steps: epoch=5 step=785 loss=8.2756
236
+ 2026-03-29 10:31:14,417 [INFO] new_opacus_codex.train_steps: epoch=5 step=790 loss=8.1827
237
+ 2026-03-29 10:31:28,435 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=790 eval_loss=7.6825 duration_sec=13.92
238
+ 2026-03-29 10:32:12,898 [INFO] new_opacus_codex.train_steps: epoch=5 step=795 loss=8.2289
239
+ 2026-03-29 10:32:55,881 [INFO] new_opacus_codex.train_steps: epoch=5 step=800 loss=8.3298
240
+ 2026-03-29 10:33:09,908 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=800 eval_loss=7.6768 duration_sec=14.02
241
+ 2026-03-29 10:33:53,775 [INFO] new_opacus_codex.train_steps: epoch=5 step=805 loss=8.3529
242
+ 2026-03-29 10:34:37,679 [INFO] new_opacus_codex.train_steps: epoch=5 step=810 loss=8.2407
243
+ 2026-03-29 10:34:51,776 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=810 eval_loss=7.6709 duration_sec=14.05
244
+ 2026-03-29 10:35:35,218 [INFO] new_opacus_codex.train_steps: epoch=5 step=815 loss=8.1297
245
+ 2026-03-29 10:36:19,219 [INFO] new_opacus_codex.train_steps: epoch=5 step=820 loss=7.9389
246
+ 2026-03-29 10:36:33,226 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=820 eval_loss=7.6660 duration_sec=13.99
247
+ 2026-03-29 10:37:16,144 [INFO] new_opacus_codex.train_steps: epoch=5 step=825 loss=7.9210
248
+ 2026-03-29 10:38:00,307 [INFO] new_opacus_codex.train_steps: epoch=5 step=830 loss=7.9761
249
+ 2026-03-29 10:38:14,348 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=830 eval_loss=7.6614 duration_sec=13.94
250
+ 2026-03-29 10:39:00,219 [INFO] new_opacus_codex.train_steps: epoch=5 step=835 loss=7.8368
251
+ 2026-03-29 10:39:43,725 [INFO] new_opacus_codex.train_steps: epoch=5 step=840 loss=7.9920
252
+ 2026-03-29 10:39:57,769 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=840 eval_loss=7.6565 duration_sec=14.03
253
+ 2026-03-29 10:40:41,596 [INFO] new_opacus_codex.train_steps: epoch=5 step=845 loss=8.0642
254
+ 2026-03-29 10:41:24,499 [INFO] new_opacus_codex.train_steps: epoch=5 step=850 loss=7.9635
255
+ 2026-03-29 10:41:38,504 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=850 eval_loss=7.6520 duration_sec=13.89
256
+ 2026-03-29 10:42:23,110 [INFO] new_opacus_codex.train_steps: epoch=5 step=855 loss=8.0653
257
+ 2026-03-29 10:43:06,624 [INFO] new_opacus_codex.train_steps: epoch=5 step=860 loss=7.9602
258
+ 2026-03-29 10:43:20,763 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=860 eval_loss=7.6483 duration_sec=14.07
259
+ 2026-03-29 10:44:04,983 [INFO] new_opacus_codex.train_steps: epoch=5 step=865 loss=7.8977
260
+ 2026-03-29 10:44:48,918 [INFO] new_opacus_codex.train_steps: epoch=5 step=870 loss=8.0040
261
+ 2026-03-29 10:45:02,930 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=870 eval_loss=7.6451 duration_sec=13.94
262
+ 2026-03-29 10:45:47,727 [INFO] new_opacus_codex.train_steps: epoch=5 step=875 loss=8.0359
263
+ 2026-03-29 10:46:32,473 [INFO] new_opacus_codex.train_steps: epoch=5 step=880 loss=8.0456
264
+ 2026-03-29 10:46:46,451 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=880 eval_loss=7.6421 duration_sec=13.96
265
+ 2026-03-29 10:47:29,866 [INFO] new_opacus_codex.train_steps: epoch=5 step=885 loss=8.0852
266
+ 2026-03-29 10:48:13,252 [INFO] new_opacus_codex.train_steps: epoch=5 step=890 loss=8.0005
267
+ 2026-03-29 10:48:27,292 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=890 eval_loss=7.6385 duration_sec=13.83
268
+ 2026-03-29 10:49:12,166 [INFO] new_opacus_codex.train_steps: epoch=5 step=895 loss=7.9273
269
+ 2026-03-29 10:49:55,942 [INFO] new_opacus_codex.train_steps: epoch=5 step=900 loss=8.0046
270
+ 2026-03-29 10:50:09,919 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=900 eval_loss=7.6345 duration_sec=13.80
271
+ 2026-03-29 10:50:53,657 [INFO] new_opacus_codex.train_steps: epoch=5 step=905 loss=8.2489
272
+ 2026-03-29 10:51:37,971 [INFO] new_opacus_codex.train_steps: epoch=5 step=910 loss=8.1815
273
+ 2026-03-29 10:51:51,873 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=910 eval_loss=7.6311 duration_sec=13.83
274
+ 2026-03-29 10:52:36,700 [INFO] new_opacus_codex.train_steps: epoch=5 step=915 loss=8.0412
275
+ 2026-03-29 10:53:12,172 [INFO] new_opacus_codex.train_steps: epoch=5 step=920 loss=8.2607
276
+ 2026-03-29 10:53:26,175 [INFO] new_opacus_codex.train_steps: eval event=eval_step epoch=5 step=920 eval_loss=7.6272 duration_sec=13.95
qwen1.5-moe-a2.7b/dp3_attn_v2/adapter/adapter_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "Qwen/Qwen1.5-MoE-A2.7B",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": true,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": [
25
+ "lm_head",
26
+ "embed_tokens"
27
+ ],
28
+ "peft_type": "LORA",
29
+ "peft_version": "0.18.1",
30
+ "qalora_group_size": 16,
31
+ "r": 16,
32
+ "rank_pattern": {},
33
+ "revision": null,
34
+ "target_modules": [
35
+ "v_proj",
36
+ "k_proj",
37
+ "q_proj",
38
+ "o_proj"
39
+ ],
40
+ "target_parameters": null,
41
+ "task_type": "CAUSAL_LM",
42
+ "trainable_token_indices": null,
43
+ "use_dora": false,
44
+ "use_qalora": false,
45
+ "use_rslora": false
46
+ }
qwen1.5-moe-a2.7b/dp3_attn_v2/tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff