Hugging Face's logo

NJUDeepEngine
/
MeteoRA-llama2-13b

Safetensors
Model card Files Community
ParagonLight commited on
Commit
b1775c3
·
1 Parent(s): a12f5b5

upload loras

Browse files
Files changed (42) hide show
  1. llama2_13b_peft/logical_deduction/README.md +73 -0
  2. llama2_13b_peft/logical_deduction/adapter_config.json +33 -0
  3. llama2_13b_peft/logical_deduction/adapter_model.safetensors +3 -0
  4. llama2_13b_peft/logical_deduction/all_results.json +11 -0
  5. llama2_13b_peft/logical_deduction/eval_results.json +7 -0
  6. llama2_13b_peft/logical_deduction/special_tokens_map.json +24 -0
  7. llama2_13b_peft/logical_deduction/tokenizer.model +3 -0
  8. llama2_13b_peft/logical_deduction/tokenizer_config.json +45 -0
  9. llama2_13b_peft/logical_deduction/train_results.json +7 -0
  10. llama2_13b_peft/logical_deduction/trainer_log.jsonl +53 -0
  11. llama2_13b_peft/logical_deduction/trainer_state.json +395 -0
  12. llama2_13b_peft/logical_deduction/training_args.bin +3 -0
  13. llama2_13b_peft/logical_deduction/training_eval_loss.png +0 -0
  14. llama2_13b_peft/logical_deduction/training_loss.png +0 -0
  15. llama2_13b_peft/news_commentary_es/README.md +84 -0
  16. llama2_13b_peft/news_commentary_es/adapter_config.json +33 -0
  17. llama2_13b_peft/news_commentary_es/adapter_model.safetensors +3 -0
  18. llama2_13b_peft/news_commentary_es/all_results.json +11 -0
  19. llama2_13b_peft/news_commentary_es/eval_results.json +7 -0
  20. llama2_13b_peft/news_commentary_es/special_tokens_map.json +24 -0
  21. llama2_13b_peft/news_commentary_es/tokenizer.model +3 -0
  22. llama2_13b_peft/news_commentary_es/tokenizer_config.json +45 -0
  23. llama2_13b_peft/news_commentary_es/train_results.json +7 -0
  24. llama2_13b_peft/news_commentary_es/trainer_log.jsonl +0 -0
  25. llama2_13b_peft/news_commentary_es/trainer_state.json +2968 -0
  26. llama2_13b_peft/news_commentary_es/training_args.bin +3 -0
  27. llama2_13b_peft/news_commentary_es/training_eval_loss.png +0 -0
  28. llama2_13b_peft/news_commentary_es/training_loss.png +0 -0
  29. llama2_13b_peft/object_counting/README.md +68 -0
  30. llama2_13b_peft/object_counting/adapter_config.json +34 -0
  31. llama2_13b_peft/object_counting/adapter_model.safetensors +3 -0
  32. llama2_13b_peft/object_counting/all_results.json +12 -0
  33. llama2_13b_peft/object_counting/eval_results.json +7 -0
  34. llama2_13b_peft/object_counting/special_tokens_map.json +24 -0
  35. llama2_13b_peft/object_counting/tokenizer.model +3 -0
  36. llama2_13b_peft/object_counting/tokenizer_config.json +45 -0
  37. llama2_13b_peft/object_counting/train_results.json +8 -0
  38. llama2_13b_peft/object_counting/trainer_log.jsonl +79 -0
  39. llama2_13b_peft/object_counting/trainer_state.json +576 -0
  40. llama2_13b_peft/object_counting/training_args.bin +3 -0
  41. llama2_13b_peft/object_counting/training_eval_loss.png +0 -0
  42. llama2_13b_peft/object_counting/training_loss.png +0 -0
llama2_13b_peft/logical_deduction/README.md ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ library_name: peft
4
+ tags:
5
+ - llama-factory
6
+ - lora
7
+ - generated_from_trainer
8
+ base_model: /data1/model/llama2/meta-llama/Llama2-13b
9
+ model-index:
10
+ - name: logical_deduction_no_sys
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # logical_deduction_no_sys
18
+
19
+ This model is a fine-tuned version of [/data1/model/llama2/meta-llama/Llama2-13b](https://huggingface.co//data1/model/llama2/meta-llama/Llama2-13b) on the logical_deduction_no_sys dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.1318
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0001
41
+ - train_batch_size: 8
42
+ - eval_batch_size: 8
43
+ - seed: 42
44
+ - distributed_type: multi-GPU
45
+ - num_devices: 3
46
+ - total_train_batch_size: 24
47
+ - total_eval_batch_size: 24
48
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_steps: 20
51
+ - num_epochs: 10.0
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:-----:|:----:|:---------------:|
57
+ | 0.121 | 1.16 | 50 | 0.1113 |
58
+ | 0.0574 | 2.33 | 100 | 0.0727 |
59
+ | 0.0405 | 3.49 | 150 | 0.0608 |
60
+ | 0.0319 | 4.65 | 200 | 0.0899 |
61
+ | 0.0236 | 5.81 | 250 | 0.0640 |
62
+ | 0.0152 | 6.98 | 300 | 0.0996 |
63
+ | 0.0053 | 8.14 | 350 | 0.1150 |
64
+ | 0.0021 | 9.3 | 400 | 0.1312 |
65
+
66
+
67
+ ### Framework versions
68
+
69
+ - PEFT 0.9.0
70
+ - Transformers 4.38.2
71
+ - Pytorch 2.2.1
72
+ - Datasets 2.18.0
73
+ - Tokenizers 0.15.2
llama2_13b_peft/logical_deduction/adapter_config.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data1/model/llama2/meta-llama/Llama2-13b",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layers_pattern": null,
10
+ "layers_to_transform": null,
11
+ "loftq_config": {},
12
+ "lora_alpha": 16,
13
+ "lora_dropout": 0.0,
14
+ "megatron_config": null,
15
+ "megatron_core": "megatron.core",
16
+ "modules_to_save": null,
17
+ "peft_type": "LORA",
18
+ "r": 8,
19
+ "rank_pattern": {},
20
+ "revision": null,
21
+ "target_modules": [
22
+ "gate_proj",
23
+ "k_proj",
24
+ "q_proj",
25
+ "down_proj",
26
+ "up_proj",
27
+ "o_proj",
28
+ "v_proj"
29
+ ],
30
+ "task_type": "CAUSAL_LM",
31
+ "use_dora": false,
32
+ "use_rslora": false
33
+ }
llama2_13b_peft/logical_deduction/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52d94d189237fab0fa3b831e5e3d81d2f34e9070ecf232ca2d1a17d2b0faee80
3
+ size 125248064
llama2_13b_peft/logical_deduction/all_results.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_loss": 0.1317996233701706,
4
+ "eval_runtime": 2.219,
5
+ "eval_samples_per_second": 81.118,
6
+ "eval_steps_per_second": 3.605,
7
+ "train_loss": 0.0926927653899373,
8
+ "train_runtime": 420.9269,
9
+ "train_samples_per_second": 24.232,
10
+ "train_steps_per_second": 1.022
11
+ }
llama2_13b_peft/logical_deduction/eval_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_loss": 0.1317996233701706,
4
+ "eval_runtime": 2.219,
5
+ "eval_samples_per_second": 81.118,
6
+ "eval_steps_per_second": 3.605
7
+ }
llama2_13b_peft/logical_deduction/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
llama2_13b_peft/logical_deduction/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
llama2_13b_peft/logical_deduction/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content + '\\n' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }
llama2_13b_peft/logical_deduction/train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "train_loss": 0.0926927653899373,
4
+ "train_runtime": 420.9269,
5
+ "train_samples_per_second": 24.232,
6
+ "train_steps_per_second": 1.022
7
+ }
llama2_13b_peft/logical_deduction/trainer_log.jsonl ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"current_steps": 10, "total_steps": 430, "loss": 1.8041, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5e-05, "epoch": 0.23, "percentage": 2.33, "elapsed_time": "0:00:10", "remaining_time": "0:07:22"}
2
+ {"current_steps": 20, "total_steps": 430, "loss": 0.5814, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 0.0001, "epoch": 0.47, "percentage": 4.65, "elapsed_time": "0:00:19", "remaining_time": "0:06:42"}
3
+ {"current_steps": 30, "total_steps": 430, "loss": 0.1611, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.985329005918702e-05, "epoch": 0.7, "percentage": 6.98, "elapsed_time": "0:00:28", "remaining_time": "0:06:21"}
4
+ {"current_steps": 40, "total_steps": 430, "loss": 0.1289, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.941402118901744e-05, "epoch": 0.93, "percentage": 9.3, "elapsed_time": "0:00:37", "remaining_time": "0:06:09"}
5
+ {"current_steps": 50, "total_steps": 430, "loss": 0.121, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.868477119388896e-05, "epoch": 1.16, "percentage": 11.63, "elapsed_time": "0:00:47", "remaining_time": "0:05:58"}
6
+ {"current_steps": 50, "total_steps": 430, "loss": null, "eval_loss": 0.11131427437067032, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 1.16, "percentage": 11.63, "elapsed_time": "0:00:47", "remaining_time": "0:05:58"}
7
+ {"current_steps": 60, "total_steps": 430, "loss": 0.1046, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.766981960274653e-05, "epoch": 1.4, "percentage": 13.95, "elapsed_time": "0:00:58", "remaining_time": "0:06:01"}
8
+ {"current_steps": 70, "total_steps": 430, "loss": 0.0922, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.637512255510475e-05, "epoch": 1.63, "percentage": 16.28, "elapsed_time": "0:01:07", "remaining_time": "0:05:49"}
9
+ {"current_steps": 80, "total_steps": 430, "loss": 0.0957, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.480827784805278e-05, "epoch": 1.86, "percentage": 18.6, "elapsed_time": "0:01:16", "remaining_time": "0:05:36"}
10
+ {"current_steps": 90, "total_steps": 430, "loss": 0.0751, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.297848034936006e-05, "epoch": 2.09, "percentage": 20.93, "elapsed_time": "0:01:26", "remaining_time": "0:05:24"}
11
+ {"current_steps": 100, "total_steps": 430, "loss": 0.0574, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.089646803833589e-05, "epoch": 2.33, "percentage": 23.26, "elapsed_time": "0:01:35", "remaining_time": "0:05:14"}
12
+ {"current_steps": 100, "total_steps": 430, "loss": null, "eval_loss": 0.07272274047136307, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 2.33, "percentage": 23.26, "elapsed_time": "0:01:35", "remaining_time": "0:05:14"}
13
+ {"current_steps": 110, "total_steps": 430, "loss": 0.0602, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.857445899109715e-05, "epoch": 2.56, "percentage": 25.58, "elapsed_time": "0:01:46", "remaining_time": "0:05:10"}
14
+ {"current_steps": 120, "total_steps": 430, "loss": 0.069, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.602607968003935e-05, "epoch": 2.79, "percentage": 27.91, "elapsed_time": "0:01:56", "remaining_time": "0:05:00"}
15
+ {"current_steps": 130, "total_steps": 430, "loss": 0.0711, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.326628500827826e-05, "epoch": 3.02, "percentage": 30.23, "elapsed_time": "0:02:05", "remaining_time": "0:04:49"}
16
+ {"current_steps": 140, "total_steps": 430, "loss": 0.0415, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.03112705483319e-05, "epoch": 3.26, "percentage": 32.56, "elapsed_time": "0:02:14", "remaining_time": "0:04:38"}
17
+ {"current_steps": 150, "total_steps": 430, "loss": 0.0405, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.717837750006106e-05, "epoch": 3.49, "percentage": 34.88, "elapsed_time": "0:02:23", "remaining_time": "0:04:28"}
18
+ {"current_steps": 150, "total_steps": 430, "loss": null, "eval_loss": 0.060838837176561356, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 3.49, "percentage": 34.88, "elapsed_time": "0:02:23", "remaining_time": "0:04:28"}
19
+ {"current_steps": 160, "total_steps": 430, "loss": 0.0453, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.388599092561315e-05, "epoch": 3.72, "percentage": 37.21, "elapsed_time": "0:02:35", "remaining_time": "0:04:22"}
20
+ {"current_steps": 170, "total_steps": 430, "loss": 0.0536, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.045343185856701e-05, "epoch": 3.95, "percentage": 39.53, "elapsed_time": "0:02:44", "remaining_time": "0:04:11"}
21
+ {"current_steps": 180, "total_steps": 430, "loss": 0.0278, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 6.690084392042513e-05, "epoch": 4.19, "percentage": 41.86, "elapsed_time": "0:02:53", "remaining_time": "0:04:01"}
22
+ {"current_steps": 190, "total_steps": 430, "loss": 0.034, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 6.32490751098331e-05, "epoch": 4.42, "percentage": 44.19, "elapsed_time": "0:03:03", "remaining_time": "0:03:51"}
23
+ {"current_steps": 200, "total_steps": 430, "loss": 0.0319, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.951955545823342e-05, "epoch": 4.65, "percentage": 46.51, "elapsed_time": "0:03:12", "remaining_time": "0:03:41"}
24
+ {"current_steps": 200, "total_steps": 430, "loss": null, "eval_loss": 0.08988692611455917, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 4.65, "percentage": 46.51, "elapsed_time": "0:03:12", "remaining_time": "0:03:41"}
25
+ {"current_steps": 210, "total_steps": 430, "loss": 0.0376, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.573417126992003e-05, "epoch": 4.88, "percentage": 48.84, "elapsed_time": "0:03:23", "remaining_time": "0:03:33"}
26
+ {"current_steps": 220, "total_steps": 430, "loss": 0.0308, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.191513668450178e-05, "epoch": 5.12, "percentage": 51.16, "elapsed_time": "0:03:32", "remaining_time": "0:03:23"}
27
+ {"current_steps": 230, "total_steps": 430, "loss": 0.0284, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.8084863315498234e-05, "epoch": 5.35, "percentage": 53.49, "elapsed_time": "0:03:42", "remaining_time": "0:03:13"}
28
+ {"current_steps": 240, "total_steps": 430, "loss": 0.0227, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.4265828730079987e-05, "epoch": 5.58, "percentage": 55.81, "elapsed_time": "0:03:51", "remaining_time": "0:03:03"}
29
+ {"current_steps": 250, "total_steps": 430, "loss": 0.0236, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.0480444541766576e-05, "epoch": 5.81, "percentage": 58.14, "elapsed_time": "0:04:00", "remaining_time": "0:02:53"}
30
+ {"current_steps": 250, "total_steps": 430, "loss": null, "eval_loss": 0.06400923430919647, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 5.81, "percentage": 58.14, "elapsed_time": "0:04:00", "remaining_time": "0:02:53"}
31
+ {"current_steps": 260, "total_steps": 430, "loss": 0.0255, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.675092489016693e-05, "epoch": 6.05, "percentage": 60.47, "elapsed_time": "0:04:12", "remaining_time": "0:02:44"}
32
+ {"current_steps": 270, "total_steps": 430, "loss": 0.0177, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.309915607957487e-05, "epoch": 6.28, "percentage": 62.79, "elapsed_time": "0:04:21", "remaining_time": "0:02:34"}
33
+ {"current_steps": 280, "total_steps": 430, "loss": 0.0118, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.9546568141433006e-05, "epoch": 6.51, "percentage": 65.12, "elapsed_time": "0:04:30", "remaining_time": "0:02:24"}
34
+ {"current_steps": 290, "total_steps": 430, "loss": 0.0128, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.6114009074386846e-05, "epoch": 6.74, "percentage": 67.44, "elapsed_time": "0:04:39", "remaining_time": "0:02:15"}
35
+ {"current_steps": 300, "total_steps": 430, "loss": 0.0152, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.282162249993895e-05, "epoch": 6.98, "percentage": 69.77, "elapsed_time": "0:04:48", "remaining_time": "0:02:05"}
36
+ {"current_steps": 300, "total_steps": 430, "loss": null, "eval_loss": 0.09959390014410019, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 6.98, "percentage": 69.77, "elapsed_time": "0:04:48", "remaining_time": "0:02:05"}
37
+ {"current_steps": 310, "total_steps": 430, "loss": 0.007, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.9688729451668114e-05, "epoch": 7.21, "percentage": 72.09, "elapsed_time": "0:05:00", "remaining_time": "0:01:56"}
38
+ {"current_steps": 320, "total_steps": 430, "loss": 0.0107, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.673371499172174e-05, "epoch": 7.44, "percentage": 74.42, "elapsed_time": "0:05:09", "remaining_time": "0:01:46"}
39
+ {"current_steps": 330, "total_steps": 430, "loss": 0.0055, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.3973920319960655e-05, "epoch": 7.67, "percentage": 76.74, "elapsed_time": "0:05:18", "remaining_time": "0:01:36"}
40
+ {"current_steps": 340, "total_steps": 430, "loss": 0.0116, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.1425541008902851e-05, "epoch": 7.91, "percentage": 79.07, "elapsed_time": "0:05:28", "remaining_time": "0:01:26"}
41
+ {"current_steps": 350, "total_steps": 430, "loss": 0.0053, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.103531961664118e-06, "epoch": 8.14, "percentage": 81.4, "elapsed_time": "0:05:37", "remaining_time": "0:01:17"}
42
+ {"current_steps": 350, "total_steps": 430, "loss": null, "eval_loss": 0.11503978818655014, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 8.14, "percentage": 81.4, "elapsed_time": "0:05:37", "remaining_time": "0:01:17"}
43
+ {"current_steps": 360, "total_steps": 430, "loss": 0.0037, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.0215196506399515e-06, "epoch": 8.37, "percentage": 83.72, "elapsed_time": "0:05:48", "remaining_time": "0:01:07"}
44
+ {"current_steps": 370, "total_steps": 430, "loss": 0.0034, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.191722151947226e-06, "epoch": 8.6, "percentage": 86.05, "elapsed_time": "0:05:57", "remaining_time": "0:00:58"}
45
+ {"current_steps": 380, "total_steps": 430, "loss": 0.004, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.6248774448952695e-06, "epoch": 8.84, "percentage": 88.37, "elapsed_time": "0:06:07", "remaining_time": "0:00:48"}
46
+ {"current_steps": 390, "total_steps": 430, "loss": 0.0032, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.330180397253473e-06, "epoch": 9.07, "percentage": 90.7, "elapsed_time": "0:06:16", "remaining_time": "0:00:38"}
47
+ {"current_steps": 400, "total_steps": 430, "loss": 0.0021, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.3152288061110518e-06, "epoch": 9.3, "percentage": 93.02, "elapsed_time": "0:06:25", "remaining_time": "0:00:28"}
48
+ {"current_steps": 400, "total_steps": 430, "loss": null, "eval_loss": 0.13117723166942596, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 9.3, "percentage": 93.02, "elapsed_time": "0:06:25", "remaining_time": "0:00:28"}
49
+ {"current_steps": 410, "total_steps": 430, "loss": 0.0026, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.859788109825793e-07, "epoch": 9.53, "percentage": 95.35, "elapsed_time": "0:06:36", "remaining_time": "0:00:19"}
50
+ {"current_steps": 420, "total_steps": 430, "loss": 0.0027, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.4670994081297795e-07, "epoch": 9.77, "percentage": 97.67, "elapsed_time": "0:06:46", "remaining_time": "0:00:09"}
51
+ {"current_steps": 430, "total_steps": 430, "loss": 0.0016, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 0.0, "epoch": 10.0, "percentage": 100.0, "elapsed_time": "0:06:55", "remaining_time": "0:00:00"}
52
+ {"current_steps": 430, "total_steps": 430, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 10.0, "percentage": 100.0, "elapsed_time": "0:06:55", "remaining_time": "0:00:00"}
53
+ {"current_steps": 8, "total_steps": 8, "loss": null, "eval_loss": 0.1317996233701706, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 10.0, "percentage": 100.0, "elapsed_time": "0:06:57", "remaining_time": "0:00:00"}
llama2_13b_peft/logical_deduction/trainer_state.json ADDED
@@ -0,0 +1,395 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 10.0,
5
+ "eval_steps": 50,
6
+ "global_step": 430,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.23,
13
+ "grad_norm": 2.765148401260376,
14
+ "learning_rate": 5e-05,
15
+ "loss": 1.8041,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.47,
20
+ "grad_norm": 0.46828070282936096,
21
+ "learning_rate": 0.0001,
22
+ "loss": 0.5814,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.7,
27
+ "grad_norm": 0.7514229416847229,
28
+ "learning_rate": 9.985329005918702e-05,
29
+ "loss": 0.1611,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.93,
34
+ "grad_norm": 0.5662431120872498,
35
+ "learning_rate": 9.941402118901744e-05,
36
+ "loss": 0.1289,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 1.16,
41
+ "grad_norm": 0.3684688210487366,
42
+ "learning_rate": 9.868477119388896e-05,
43
+ "loss": 0.121,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 1.16,
48
+ "eval_loss": 0.11131427437067032,
49
+ "eval_runtime": 2.2112,
50
+ "eval_samples_per_second": 81.405,
51
+ "eval_steps_per_second": 3.618,
52
+ "step": 50
53
+ },
54
+ {
55
+ "epoch": 1.4,
56
+ "grad_norm": 0.2601034939289093,
57
+ "learning_rate": 9.766981960274653e-05,
58
+ "loss": 0.1046,
59
+ "step": 60
60
+ },
61
+ {
62
+ "epoch": 1.63,
63
+ "grad_norm": 0.7295863628387451,
64
+ "learning_rate": 9.637512255510475e-05,
65
+ "loss": 0.0922,
66
+ "step": 70
67
+ },
68
+ {
69
+ "epoch": 1.86,
70
+ "grad_norm": 0.3247099220752716,
71
+ "learning_rate": 9.480827784805278e-05,
72
+ "loss": 0.0957,
73
+ "step": 80
74
+ },
75
+ {
76
+ "epoch": 2.09,
77
+ "grad_norm": 0.4998391568660736,
78
+ "learning_rate": 9.297848034936006e-05,
79
+ "loss": 0.0751,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 2.33,
84
+ "grad_norm": 0.4482787549495697,
85
+ "learning_rate": 9.089646803833589e-05,
86
+ "loss": 0.0574,
87
+ "step": 100
88
+ },
89
+ {
90
+ "epoch": 2.33,
91
+ "eval_loss": 0.07272274047136307,
92
+ "eval_runtime": 2.223,
93
+ "eval_samples_per_second": 80.972,
94
+ "eval_steps_per_second": 3.599,
95
+ "step": 100
96
+ },
97
+ {
98
+ "epoch": 2.56,
99
+ "grad_norm": 0.4341526925563812,
100
+ "learning_rate": 8.857445899109715e-05,
101
+ "loss": 0.0602,
102
+ "step": 110
103
+ },
104
+ {
105
+ "epoch": 2.79,
106
+ "grad_norm": 0.33612173795700073,
107
+ "learning_rate": 8.602607968003935e-05,
108
+ "loss": 0.069,
109
+ "step": 120
110
+ },
111
+ {
112
+ "epoch": 3.02,
113
+ "grad_norm": 0.21825428307056427,
114
+ "learning_rate": 8.326628500827826e-05,
115
+ "loss": 0.0711,
116
+ "step": 130
117
+ },
118
+ {
119
+ "epoch": 3.26,
120
+ "grad_norm": 0.5240222811698914,
121
+ "learning_rate": 8.03112705483319e-05,
122
+ "loss": 0.0415,
123
+ "step": 140
124
+ },
125
+ {
126
+ "epoch": 3.49,
127
+ "grad_norm": 0.7221460342407227,
128
+ "learning_rate": 7.717837750006106e-05,
129
+ "loss": 0.0405,
130
+ "step": 150
131
+ },
132
+ {
133
+ "epoch": 3.49,
134
+ "eval_loss": 0.060838837176561356,
135
+ "eval_runtime": 2.2271,
136
+ "eval_samples_per_second": 80.821,
137
+ "eval_steps_per_second": 3.592,
138
+ "step": 150
139
+ },
140
+ {
141
+ "epoch": 3.72,
142
+ "grad_norm": 0.26117128133773804,
143
+ "learning_rate": 7.388599092561315e-05,
144
+ "loss": 0.0453,
145
+ "step": 160
146
+ },
147
+ {
148
+ "epoch": 3.95,
149
+ "grad_norm": 0.41725072264671326,
150
+ "learning_rate": 7.045343185856701e-05,
151
+ "loss": 0.0536,
152
+ "step": 170
153
+ },
154
+ {
155
+ "epoch": 4.19,
156
+ "grad_norm": 0.3449525833129883,
157
+ "learning_rate": 6.690084392042513e-05,
158
+ "loss": 0.0278,
159
+ "step": 180
160
+ },
161
+ {
162
+ "epoch": 4.42,
163
+ "grad_norm": 0.9864999055862427,
164
+ "learning_rate": 6.32490751098331e-05,
165
+ "loss": 0.034,
166
+ "step": 190
167
+ },
168
+ {
169
+ "epoch": 4.65,
170
+ "grad_norm": 0.25210413336753845,
171
+ "learning_rate": 5.951955545823342e-05,
172
+ "loss": 0.0319,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 4.65,
177
+ "eval_loss": 0.08988692611455917,
178
+ "eval_runtime": 2.2211,
179
+ "eval_samples_per_second": 81.039,
180
+ "eval_steps_per_second": 3.602,
181
+ "step": 200
182
+ },
183
+ {
184
+ "epoch": 4.88,
185
+ "grad_norm": 0.32543689012527466,
186
+ "learning_rate": 5.573417126992003e-05,
187
+ "loss": 0.0376,
188
+ "step": 210
189
+ },
190
+ {
191
+ "epoch": 5.12,
192
+ "grad_norm": 0.2424646019935608,
193
+ "learning_rate": 5.191513668450178e-05,
194
+ "loss": 0.0308,
195
+ "step": 220
196
+ },
197
+ {
198
+ "epoch": 5.35,
199
+ "grad_norm": 0.34920650720596313,
200
+ "learning_rate": 4.8084863315498234e-05,
201
+ "loss": 0.0284,
202
+ "step": 230
203
+ },
204
+ {
205
+ "epoch": 5.58,
206
+ "grad_norm": 0.4021368622779846,
207
+ "learning_rate": 4.4265828730079987e-05,
208
+ "loss": 0.0227,
209
+ "step": 240
210
+ },
211
+ {
212
+ "epoch": 5.81,
213
+ "grad_norm": 0.8135946989059448,
214
+ "learning_rate": 4.0480444541766576e-05,
215
+ "loss": 0.0236,
216
+ "step": 250
217
+ },
218
+ {
219
+ "epoch": 5.81,
220
+ "eval_loss": 0.06400923430919647,
221
+ "eval_runtime": 2.2241,
222
+ "eval_samples_per_second": 80.932,
223
+ "eval_steps_per_second": 3.597,
224
+ "step": 250
225
+ },
226
+ {
227
+ "epoch": 6.05,
228
+ "grad_norm": 0.37095046043395996,
229
+ "learning_rate": 3.675092489016693e-05,
230
+ "loss": 0.0255,
231
+ "step": 260
232
+ },
233
+ {
234
+ "epoch": 6.28,
235
+ "grad_norm": 0.18603980541229248,
236
+ "learning_rate": 3.309915607957487e-05,
237
+ "loss": 0.0177,
238
+ "step": 270
239
+ },
240
+ {
241
+ "epoch": 6.51,
242
+ "grad_norm": 0.4936978220939636,
243
+ "learning_rate": 2.9546568141433006e-05,
244
+ "loss": 0.0118,
245
+ "step": 280
246
+ },
247
+ {
248
+ "epoch": 6.74,
249
+ "grad_norm": 0.4769618809223175,
250
+ "learning_rate": 2.6114009074386846e-05,
251
+ "loss": 0.0128,
252
+ "step": 290
253
+ },
254
+ {
255
+ "epoch": 6.98,
256
+ "grad_norm": 0.4464145004749298,
257
+ "learning_rate": 2.282162249993895e-05,
258
+ "loss": 0.0152,
259
+ "step": 300
260
+ },
261
+ {
262
+ "epoch": 6.98,
263
+ "eval_loss": 0.09959390014410019,
264
+ "eval_runtime": 2.2242,
265
+ "eval_samples_per_second": 80.926,
266
+ "eval_steps_per_second": 3.597,
267
+ "step": 300
268
+ },
269
+ {
270
+ "epoch": 7.21,
271
+ "grad_norm": 0.12924155592918396,
272
+ "learning_rate": 1.9688729451668114e-05,
273
+ "loss": 0.007,
274
+ "step": 310
275
+ },
276
+ {
277
+ "epoch": 7.44,
278
+ "grad_norm": 0.4371550381183624,
279
+ "learning_rate": 1.673371499172174e-05,
280
+ "loss": 0.0107,
281
+ "step": 320
282
+ },
283
+ {
284
+ "epoch": 7.67,
285
+ "grad_norm": 0.23555269837379456,
286
+ "learning_rate": 1.3973920319960655e-05,
287
+ "loss": 0.0055,
288
+ "step": 330
289
+ },
290
+ {
291
+ "epoch": 7.91,
292
+ "grad_norm": 0.517039954662323,
293
+ "learning_rate": 1.1425541008902851e-05,
294
+ "loss": 0.0116,
295
+ "step": 340
296
+ },
297
+ {
298
+ "epoch": 8.14,
299
+ "grad_norm": 0.026530884206295013,
300
+ "learning_rate": 9.103531961664118e-06,
301
+ "loss": 0.0053,
302
+ "step": 350
303
+ },
304
+ {
305
+ "epoch": 8.14,
306
+ "eval_loss": 0.11503978818655014,
307
+ "eval_runtime": 2.2223,
308
+ "eval_samples_per_second": 80.998,
309
+ "eval_steps_per_second": 3.6,
310
+ "step": 350
311
+ },
312
+ {
313
+ "epoch": 8.37,
314
+ "grad_norm": 0.2660503685474396,
315
+ "learning_rate": 7.0215196506399515e-06,
316
+ "loss": 0.0037,
317
+ "step": 360
318
+ },
319
+ {
320
+ "epoch": 8.6,
321
+ "grad_norm": 0.5767481327056885,
322
+ "learning_rate": 5.191722151947226e-06,
323
+ "loss": 0.0034,
324
+ "step": 370
325
+ },
326
+ {
327
+ "epoch": 8.84,
328
+ "grad_norm": 0.9280475974082947,
329
+ "learning_rate": 3.6248774448952695e-06,
330
+ "loss": 0.004,
331
+ "step": 380
332
+ },
333
+ {
334
+ "epoch": 9.07,
335
+ "grad_norm": 0.0765455961227417,
336
+ "learning_rate": 2.330180397253473e-06,
337
+ "loss": 0.0032,
338
+ "step": 390
339
+ },
340
+ {
341
+ "epoch": 9.3,
342
+ "grad_norm": 0.05842304602265358,
343
+ "learning_rate": 1.3152288061110518e-06,
344
+ "loss": 0.0021,
345
+ "step": 400
346
+ },
347
+ {
348
+ "epoch": 9.3,
349
+ "eval_loss": 0.13117723166942596,
350
+ "eval_runtime": 2.2239,
351
+ "eval_samples_per_second": 80.939,
352
+ "eval_steps_per_second": 3.597,
353
+ "step": 400
354
+ },
355
+ {
356
+ "epoch": 9.53,
357
+ "grad_norm": 0.02898208238184452,
358
+ "learning_rate": 5.859788109825793e-07,
359
+ "loss": 0.0026,
360
+ "step": 410
361
+ },
362
+ {
363
+ "epoch": 9.77,
364
+ "grad_norm": 0.41006436944007874,
365
+ "learning_rate": 1.4670994081297795e-07,
366
+ "loss": 0.0027,
367
+ "step": 420
368
+ },
369
+ {
370
+ "epoch": 10.0,
371
+ "grad_norm": 0.015619861893355846,
372
+ "learning_rate": 0.0,
373
+ "loss": 0.0016,
374
+ "step": 430
375
+ },
376
+ {
377
+ "epoch": 10.0,
378
+ "step": 430,
379
+ "total_flos": 2.0536791678110925e+17,
380
+ "train_loss": 0.0926927653899373,
381
+ "train_runtime": 420.9269,
382
+ "train_samples_per_second": 24.232,
383
+ "train_steps_per_second": 1.022
384
+ }
385
+ ],
386
+ "logging_steps": 10,
387
+ "max_steps": 430,
388
+ "num_input_tokens_seen": 0,
389
+ "num_train_epochs": 10,
390
+ "save_steps": 500,
391
+ "total_flos": 2.0536791678110925e+17,
392
+ "train_batch_size": 8,
393
+ "trial_name": null,
394
+ "trial_params": null
395
+ }
llama2_13b_peft/logical_deduction/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9126db8bbe3c6a12001b7b0a43c2c93caf2551adcba85b55ad121e49d0420ac
3
+ size 5112
llama2_13b_peft/logical_deduction/training_eval_loss.png ADDED
llama2_13b_peft/logical_deduction/training_loss.png ADDED
llama2_13b_peft/news_commentary_es/README.md ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ library_name: peft
4
+ tags:
5
+ - llama-factory
6
+ - lora
7
+ - generated_from_trainer
8
+ base_model: /data1/model/llama2/meta-llama/Llama2-13b
9
+ model-index:
10
+ - name: news_commentary_es
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # news_commentary_es
18
+
19
+ This model is a fine-tuned version of [/data1/model/llama2/meta-llama/Llama2-13b](https://huggingface.co//data1/model/llama2/meta-llama/Llama2-13b) on the news_commentary_es dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.5704
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 1e-05
41
+ - train_batch_size: 16
42
+ - eval_batch_size: 16
43
+ - seed: 42
44
+ - distributed_type: multi-GPU
45
+ - num_devices: 2
46
+ - total_train_batch_size: 32
47
+ - total_eval_batch_size: 32
48
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_steps: 20
51
+ - num_epochs: 5.0
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:-----:|:----:|:---------------:|
57
+ | 0.6296 | 0.25 | 200 | 0.6344 |
58
+ | 0.6272 | 0.5 | 400 | 0.6124 |
59
+ | 0.6167 | 0.75 | 600 | 0.6020 |
60
+ | 0.5882 | 1.0 | 800 | 0.5940 |
61
+ | 0.5522 | 1.25 | 1000 | 0.5889 |
62
+ | 0.5919 | 1.51 | 1200 | 0.5836 |
63
+ | 0.6038 | 1.76 | 1400 | 0.5801 |
64
+ | 0.5882 | 2.01 | 1600 | 0.5773 |
65
+ | 0.5302 | 2.26 | 1800 | 0.5754 |
66
+ | 0.5331 | 2.51 | 2000 | 0.5732 |
67
+ | 0.5669 | 2.76 | 2200 | 0.5714 |
68
+ | 0.5551 | 3.01 | 2400 | 0.5706 |
69
+ | 0.5499 | 3.26 | 2600 | 0.5716 |
70
+ | 0.545 | 3.51 | 2800 | 0.5704 |
71
+ | 0.5253 | 3.76 | 3000 | 0.5704 |
72
+ | 0.5418 | 4.02 | 3200 | 0.5702 |
73
+ | 0.5289 | 4.27 | 3400 | 0.5704 |
74
+ | 0.4983 | 4.52 | 3600 | 0.5706 |
75
+ | 0.5371 | 4.77 | 3800 | 0.5705 |
76
+
77
+
78
+ ### Framework versions
79
+
80
+ - PEFT 0.9.0
81
+ - Transformers 4.38.2
82
+ - Pytorch 2.2.1
83
+ - Datasets 2.18.0
84
+ - Tokenizers 0.15.2
llama2_13b_peft/news_commentary_es/adapter_config.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data1/model/llama2/meta-llama/Llama2-13b",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layers_pattern": null,
10
+ "layers_to_transform": null,
11
+ "loftq_config": {},
12
+ "lora_alpha": 16,
13
+ "lora_dropout": 0.0,
14
+ "megatron_config": null,
15
+ "megatron_core": "megatron.core",
16
+ "modules_to_save": null,
17
+ "peft_type": "LORA",
18
+ "r": 8,
19
+ "rank_pattern": {},
20
+ "revision": null,
21
+ "target_modules": [
22
+ "q_proj",
23
+ "k_proj",
24
+ "gate_proj",
25
+ "down_proj",
26
+ "o_proj",
27
+ "up_proj",
28
+ "v_proj"
29
+ ],
30
+ "task_type": "CAUSAL_LM",
31
+ "use_dora": false,
32
+ "use_rslora": false
33
+ }
llama2_13b_peft/news_commentary_es/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e8a2b2fd9e200ef442df6c06579336a490267572c4c62ea475e8fd466214a8bf
3
+ size 125248064
llama2_13b_peft/news_commentary_es/all_results.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_loss": 0.5704381465911865,
4
+ "eval_runtime": 68.1349,
5
+ "eval_samples_per_second": 66.045,
6
+ "eval_steps_per_second": 2.069,
7
+ "train_loss": 0.569244503227052,
8
+ "train_runtime": 8741.9073,
9
+ "train_samples_per_second": 14.585,
10
+ "train_steps_per_second": 0.456
11
+ }
llama2_13b_peft/news_commentary_es/eval_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_loss": 0.5704381465911865,
4
+ "eval_runtime": 68.1349,
5
+ "eval_samples_per_second": 66.045,
6
+ "eval_steps_per_second": 2.069
7
+ }
llama2_13b_peft/news_commentary_es/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
llama2_13b_peft/news_commentary_es/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
llama2_13b_peft/news_commentary_es/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content + '\\n' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }
llama2_13b_peft/news_commentary_es/train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "train_loss": 0.569244503227052,
4
+ "train_runtime": 8741.9073,
5
+ "train_samples_per_second": 14.585,
6
+ "train_steps_per_second": 0.456
7
+ }
llama2_13b_peft/news_commentary_es/trainer_log.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
llama2_13b_peft/news_commentary_es/trainer_state.json ADDED
@@ -0,0 +1,2968 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.5704381465911865,
3
+ "best_model_checkpoint": "ckpt/llama2_13b_fuze30_no_sys/news_commentary_es/checkpoint-3000",
4
+ "epoch": 5.0,
5
+ "eval_steps": 200,
6
+ "global_step": 3985,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "grad_norm": 0.5029562711715698,
14
+ "learning_rate": 5e-06,
15
+ "loss": 1.2458,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.03,
20
+ "grad_norm": 0.6982043981552124,
21
+ "learning_rate": 1e-05,
22
+ "loss": 1.266,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.04,
27
+ "grad_norm": 0.7867264151573181,
28
+ "learning_rate": 9.99984305369385e-06,
29
+ "loss": 1.2072,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.05,
34
+ "grad_norm": 0.7258906364440918,
35
+ "learning_rate": 9.999372224628255e-06,
36
+ "loss": 1.0755,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.06,
41
+ "grad_norm": 0.6007847189903259,
42
+ "learning_rate": 9.998587542361168e-06,
43
+ "loss": 0.9002,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.08,
48
+ "grad_norm": 0.3854537606239319,
49
+ "learning_rate": 9.997489056153784e-06,
50
+ "loss": 0.7568,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.09,
55
+ "grad_norm": 0.3243318796157837,
56
+ "learning_rate": 9.996076834967442e-06,
57
+ "loss": 0.7157,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.1,
62
+ "grad_norm": 0.24546250700950623,
63
+ "learning_rate": 9.994350967459303e-06,
64
+ "loss": 0.6984,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.11,
69
+ "grad_norm": 0.34287428855895996,
70
+ "learning_rate": 9.99231156197678e-06,
71
+ "loss": 0.6996,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.13,
76
+ "grad_norm": 0.3073829412460327,
77
+ "learning_rate": 9.989958746550732e-06,
78
+ "loss": 0.7129,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.14,
83
+ "grad_norm": 0.2948177456855774,
84
+ "learning_rate": 9.987292668887439e-06,
85
+ "loss": 0.6977,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.15,
90
+ "grad_norm": 0.3140040934085846,
91
+ "learning_rate": 9.984313496359317e-06,
92
+ "loss": 0.6272,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.16,
97
+ "grad_norm": 0.2903355658054352,
98
+ "learning_rate": 9.981021415994413e-06,
99
+ "loss": 0.6285,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.18,
104
+ "grad_norm": 0.3696918189525604,
105
+ "learning_rate": 9.97741663446467e-06,
106
+ "loss": 0.651,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.19,
111
+ "grad_norm": 0.24330583214759827,
112
+ "learning_rate": 9.973499378072947e-06,
113
+ "loss": 0.6376,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.2,
118
+ "grad_norm": 0.26600131392478943,
119
+ "learning_rate": 9.96926989273881e-06,
120
+ "loss": 0.6402,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.21,
125
+ "grad_norm": 0.34049201011657715,
126
+ "learning_rate": 9.9647284439831e-06,
127
+ "loss": 0.6662,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.23,
132
+ "grad_norm": 0.3909376859664917,
133
+ "learning_rate": 9.959875316911261e-06,
134
+ "loss": 0.6134,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.24,
139
+ "grad_norm": 0.3953985869884491,
140
+ "learning_rate": 9.95471081619544e-06,
141
+ "loss": 0.6054,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.25,
146
+ "grad_norm": 0.42383915185928345,
147
+ "learning_rate": 9.949235266055356e-06,
148
+ "loss": 0.6296,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.25,
153
+ "eval_loss": 0.6343660950660706,
154
+ "eval_runtime": 71.7569,
155
+ "eval_samples_per_second": 62.712,
156
+ "eval_steps_per_second": 1.965,
157
+ "step": 200
158
+ },
159
+ {
160
+ "epoch": 0.26,
161
+ "grad_norm": 0.37584221363067627,
162
+ "learning_rate": 9.943449010237965e-06,
163
+ "loss": 0.6724,
164
+ "step": 210
165
+ },
166
+ {
167
+ "epoch": 0.28,
168
+ "grad_norm": 0.368352472782135,
169
+ "learning_rate": 9.937352411995851e-06,
170
+ "loss": 0.6065,
171
+ "step": 220
172
+ },
173
+ {
174
+ "epoch": 0.29,
175
+ "grad_norm": 0.30686938762664795,
176
+ "learning_rate": 9.930945854064447e-06,
177
+ "loss": 0.5998,
178
+ "step": 230
179
+ },
180
+ {
181
+ "epoch": 0.3,
182
+ "grad_norm": 0.2913750410079956,
183
+ "learning_rate": 9.924229738637992e-06,
184
+ "loss": 0.6274,
185
+ "step": 240
186
+ },
187
+ {
188
+ "epoch": 0.31,
189
+ "grad_norm": 0.30744606256484985,
190
+ "learning_rate": 9.91720448734429e-06,
191
+ "loss": 0.6442,
192
+ "step": 250
193
+ },
194
+ {
195
+ "epoch": 0.33,
196
+ "grad_norm": 0.315236359834671,
197
+ "learning_rate": 9.909870541218238e-06,
198
+ "loss": 0.6026,
199
+ "step": 260
200
+ },
201
+ {
202
+ "epoch": 0.34,
203
+ "grad_norm": 0.28183722496032715,
204
+ "learning_rate": 9.902228360674137e-06,
205
+ "loss": 0.6511,
206
+ "step": 270
207
+ },
208
+ {
209
+ "epoch": 0.35,
210
+ "grad_norm": 0.33093082904815674,
211
+ "learning_rate": 9.89427842547679e-06,
212
+ "loss": 0.6575,
213
+ "step": 280
214
+ },
215
+ {
216
+ "epoch": 0.36,
217
+ "grad_norm": 0.3711361587047577,
218
+ "learning_rate": 9.886021234711382e-06,
219
+ "loss": 0.6209,
220
+ "step": 290
221
+ },
222
+ {
223
+ "epoch": 0.38,
224
+ "grad_norm": 0.35551759600639343,
225
+ "learning_rate": 9.877457306752149e-06,
226
+ "loss": 0.6265,
227
+ "step": 300
228
+ },
229
+ {
230
+ "epoch": 0.39,
231
+ "grad_norm": 0.3064817786216736,
232
+ "learning_rate": 9.868587179229834e-06,
233
+ "loss": 0.6214,
234
+ "step": 310
235
+ },
236
+ {
237
+ "epoch": 0.4,
238
+ "grad_norm": 0.4653981924057007,
239
+ "learning_rate": 9.859411408997939e-06,
240
+ "loss": 0.6234,
241
+ "step": 320
242
+ },
243
+ {
244
+ "epoch": 0.41,
245
+ "grad_norm": 0.3303908407688141,
246
+ "learning_rate": 9.84993057209776e-06,
247
+ "loss": 0.6502,
248
+ "step": 330
249
+ },
250
+ {
251
+ "epoch": 0.43,
252
+ "grad_norm": 0.3864089548587799,
253
+ "learning_rate": 9.84014526372223e-06,
254
+ "loss": 0.6275,
255
+ "step": 340
256
+ },
257
+ {
258
+ "epoch": 0.44,
259
+ "grad_norm": 0.3393557071685791,
260
+ "learning_rate": 9.83005609817855e-06,
261
+ "loss": 0.5913,
262
+ "step": 350
263
+ },
264
+ {
265
+ "epoch": 0.45,
266
+ "grad_norm": 0.3689006567001343,
267
+ "learning_rate": 9.819663708849624e-06,
268
+ "loss": 0.5962,
269
+ "step": 360
270
+ },
271
+ {
272
+ "epoch": 0.46,
273
+ "grad_norm": 0.32422763109207153,
274
+ "learning_rate": 9.808968748154302e-06,
275
+ "loss": 0.5896,
276
+ "step": 370
277
+ },
278
+ {
279
+ "epoch": 0.48,
280
+ "grad_norm": 0.3583005964756012,
281
+ "learning_rate": 9.797971887506413e-06,
282
+ "loss": 0.617,
283
+ "step": 380
284
+ },
285
+ {
286
+ "epoch": 0.49,
287
+ "grad_norm": 0.331403911113739,
288
+ "learning_rate": 9.786673817272621e-06,
289
+ "loss": 0.5994,
290
+ "step": 390
291
+ },
292
+ {
293
+ "epoch": 0.5,
294
+ "grad_norm": 0.3698234558105469,
295
+ "learning_rate": 9.77507524672908e-06,
296
+ "loss": 0.6272,
297
+ "step": 400
298
+ },
299
+ {
300
+ "epoch": 0.5,
301
+ "eval_loss": 0.612372875213623,
302
+ "eval_runtime": 68.9986,
303
+ "eval_samples_per_second": 65.219,
304
+ "eval_steps_per_second": 2.044,
305
+ "step": 400
306
+ },
307
+ {
308
+ "epoch": 0.51,
309
+ "grad_norm": 0.3716101348400116,
310
+ "learning_rate": 9.763176904016914e-06,
311
+ "loss": 0.6246,
312
+ "step": 410
313
+ },
314
+ {
315
+ "epoch": 0.53,
316
+ "grad_norm": 0.4471350610256195,
317
+ "learning_rate": 9.750979536096495e-06,
318
+ "loss": 0.645,
319
+ "step": 420
320
+ },
321
+ {
322
+ "epoch": 0.54,
323
+ "grad_norm": 0.31366682052612305,
324
+ "learning_rate": 9.738483908700563e-06,
325
+ "loss": 0.6397,
326
+ "step": 430
327
+ },
328
+ {
329
+ "epoch": 0.55,
330
+ "grad_norm": 0.3895745575428009,
331
+ "learning_rate": 9.72569080628614e-06,
332
+ "loss": 0.6118,
333
+ "step": 440
334
+ },
335
+ {
336
+ "epoch": 0.56,
337
+ "grad_norm": 0.42779144644737244,
338
+ "learning_rate": 9.71260103198529e-06,
339
+ "loss": 0.6359,
340
+ "step": 450
341
+ },
342
+ {
343
+ "epoch": 0.58,
344
+ "grad_norm": 0.37752124667167664,
345
+ "learning_rate": 9.699215407554712e-06,
346
+ "loss": 0.5863,
347
+ "step": 460
348
+ },
349
+ {
350
+ "epoch": 0.59,
351
+ "grad_norm": 0.43578141927719116,
352
+ "learning_rate": 9.685534773324124e-06,
353
+ "loss": 0.6244,
354
+ "step": 470
355
+ },
356
+ {
357
+ "epoch": 0.6,
358
+ "grad_norm": 0.5181850790977478,
359
+ "learning_rate": 9.671559988143529e-06,
360
+ "loss": 0.5962,
361
+ "step": 480
362
+ },
363
+ {
364
+ "epoch": 0.61,
365
+ "grad_norm": 0.43225836753845215,
366
+ "learning_rate": 9.657291929329293e-06,
367
+ "loss": 0.631,
368
+ "step": 490
369
+ },
370
+ {
371
+ "epoch": 0.63,
372
+ "grad_norm": 0.36941251158714294,
373
+ "learning_rate": 9.64273149260907e-06,
374
+ "loss": 0.6398,
375
+ "step": 500
376
+ },
377
+ {
378
+ "epoch": 0.64,
379
+ "grad_norm": 0.3803509473800659,
380
+ "learning_rate": 9.627879592065558e-06,
381
+ "loss": 0.6393,
382
+ "step": 510
383
+ },
384
+ {
385
+ "epoch": 0.65,
386
+ "grad_norm": 0.3586438298225403,
387
+ "learning_rate": 9.612737160079134e-06,
388
+ "loss": 0.5944,
389
+ "step": 520
390
+ },
391
+ {
392
+ "epoch": 0.66,
393
+ "grad_norm": 0.41518551111221313,
394
+ "learning_rate": 9.597305147269303e-06,
395
+ "loss": 0.6224,
396
+ "step": 530
397
+ },
398
+ {
399
+ "epoch": 0.68,
400
+ "grad_norm": 0.3281481862068176,
401
+ "learning_rate": 9.581584522435025e-06,
402
+ "loss": 0.5853,
403
+ "step": 540
404
+ },
405
+ {
406
+ "epoch": 0.69,
407
+ "grad_norm": 0.3723869323730469,
408
+ "learning_rate": 9.565576272493902e-06,
409
+ "loss": 0.591,
410
+ "step": 550
411
+ },
412
+ {
413
+ "epoch": 0.7,
414
+ "grad_norm": 0.37543514370918274,
415
+ "learning_rate": 9.549281402420212e-06,
416
+ "loss": 0.5883,
417
+ "step": 560
418
+ },
419
+ {
420
+ "epoch": 0.72,
421
+ "grad_norm": 0.39954090118408203,
422
+ "learning_rate": 9.532700935181822e-06,
423
+ "loss": 0.5942,
424
+ "step": 570
425
+ },
426
+ {
427
+ "epoch": 0.73,
428
+ "grad_norm": 0.4060945212841034,
429
+ "learning_rate": 9.515835911675968e-06,
430
+ "loss": 0.6082,
431
+ "step": 580
432
+ },
433
+ {
434
+ "epoch": 0.74,
435
+ "grad_norm": 0.32310950756073,
436
+ "learning_rate": 9.498687390663905e-06,
437
+ "loss": 0.5836,
438
+ "step": 590
439
+ },
440
+ {
441
+ "epoch": 0.75,
442
+ "grad_norm": 0.37066149711608887,
443
+ "learning_rate": 9.481256448704445e-06,
444
+ "loss": 0.6167,
445
+ "step": 600
446
+ },
447
+ {
448
+ "epoch": 0.75,
449
+ "eval_loss": 0.6020291447639465,
450
+ "eval_runtime": 71.3713,
451
+ "eval_samples_per_second": 63.051,
452
+ "eval_steps_per_second": 1.976,
453
+ "step": 600
454
+ },
455
+ {
456
+ "epoch": 0.77,
457
+ "grad_norm": 0.38338422775268555,
458
+ "learning_rate": 9.463544180086368e-06,
459
+ "loss": 0.6211,
460
+ "step": 610
461
+ },
462
+ {
463
+ "epoch": 0.78,
464
+ "grad_norm": 0.3963073790073395,
465
+ "learning_rate": 9.445551696759732e-06,
466
+ "loss": 0.5827,
467
+ "step": 620
468
+ },
469
+ {
470
+ "epoch": 0.79,
471
+ "grad_norm": 0.4199773967266083,
472
+ "learning_rate": 9.427280128266049e-06,
473
+ "loss": 0.5782,
474
+ "step": 630
475
+ },
476
+ {
477
+ "epoch": 0.8,
478
+ "grad_norm": 0.39564579725265503,
479
+ "learning_rate": 9.4087306216674e-06,
480
+ "loss": 0.6028,
481
+ "step": 640
482
+ },
483
+ {
484
+ "epoch": 0.82,
485
+ "grad_norm": 0.43293875455856323,
486
+ "learning_rate": 9.389904341474394e-06,
487
+ "loss": 0.5791,
488
+ "step": 650
489
+ },
490
+ {
491
+ "epoch": 0.83,
492
+ "grad_norm": 0.3330543041229248,
493
+ "learning_rate": 9.370802469573088e-06,
494
+ "loss": 0.5875,
495
+ "step": 660
496
+ },
497
+ {
498
+ "epoch": 0.84,
499
+ "grad_norm": 0.4487074613571167,
500
+ "learning_rate": 9.351426205150778e-06,
501
+ "loss": 0.6148,
502
+ "step": 670
503
+ },
504
+ {
505
+ "epoch": 0.85,
506
+ "grad_norm": 0.4278540015220642,
507
+ "learning_rate": 9.331776764620711e-06,
508
+ "loss": 0.6271,
509
+ "step": 680
510
+ },
511
+ {
512
+ "epoch": 0.87,
513
+ "grad_norm": 0.4620065987110138,
514
+ "learning_rate": 9.311855381545735e-06,
515
+ "loss": 0.6439,
516
+ "step": 690
517
+ },
518
+ {
519
+ "epoch": 0.88,
520
+ "grad_norm": 0.5026012063026428,
521
+ "learning_rate": 9.291663306560843e-06,
522
+ "loss": 0.614,
523
+ "step": 700
524
+ },
525
+ {
526
+ "epoch": 0.89,
527
+ "grad_norm": 0.3847663700580597,
528
+ "learning_rate": 9.271201807294667e-06,
529
+ "loss": 0.6058,
530
+ "step": 710
531
+ },
532
+ {
533
+ "epoch": 0.9,
534
+ "grad_norm": 0.38854289054870605,
535
+ "learning_rate": 9.250472168289901e-06,
536
+ "loss": 0.5859,
537
+ "step": 720
538
+ },
539
+ {
540
+ "epoch": 0.92,
541
+ "grad_norm": 0.4080798625946045,
542
+ "learning_rate": 9.22947569092265e-06,
543
+ "loss": 0.5874,
544
+ "step": 730
545
+ },
546
+ {
547
+ "epoch": 0.93,
548
+ "grad_norm": 0.587581992149353,
549
+ "learning_rate": 9.20821369332074e-06,
550
+ "loss": 0.6077,
551
+ "step": 740
552
+ },
553
+ {
554
+ "epoch": 0.94,
555
+ "grad_norm": 0.5180934071540833,
556
+ "learning_rate": 9.186687510280968e-06,
557
+ "loss": 0.6005,
558
+ "step": 750
559
+ },
560
+ {
561
+ "epoch": 0.95,
562
+ "grad_norm": 0.4063900411128998,
563
+ "learning_rate": 9.164898493185295e-06,
564
+ "loss": 0.6007,
565
+ "step": 760
566
+ },
567
+ {
568
+ "epoch": 0.97,
569
+ "grad_norm": 0.3806501626968384,
570
+ "learning_rate": 9.142848009916025e-06,
571
+ "loss": 0.6,
572
+ "step": 770
573
+ },
574
+ {
575
+ "epoch": 0.98,
576
+ "grad_norm": 0.4454844892024994,
577
+ "learning_rate": 9.120537444769912e-06,
578
+ "loss": 0.5704,
579
+ "step": 780
580
+ },
581
+ {
582
+ "epoch": 0.99,
583
+ "grad_norm": 0.4750034511089325,
584
+ "learning_rate": 9.097968198371277e-06,
585
+ "loss": 0.626,
586
+ "step": 790
587
+ },
588
+ {
589
+ "epoch": 1.0,
590
+ "grad_norm": 0.44586676359176636,
591
+ "learning_rate": 9.075141687584056e-06,
592
+ "loss": 0.5882,
593
+ "step": 800
594
+ },
595
+ {
596
+ "epoch": 1.0,
597
+ "eval_loss": 0.5939783453941345,
598
+ "eval_runtime": 72.333,
599
+ "eval_samples_per_second": 62.212,
600
+ "eval_steps_per_second": 1.949,
601
+ "step": 800
602
+ },
603
+ {
604
+ "epoch": 1.02,
605
+ "grad_norm": 0.4351860284805298,
606
+ "learning_rate": 9.052059345422874e-06,
607
+ "loss": 0.5908,
608
+ "step": 810
609
+ },
610
+ {
611
+ "epoch": 1.03,
612
+ "grad_norm": 0.5268151760101318,
613
+ "learning_rate": 9.028722620963066e-06,
614
+ "loss": 0.6155,
615
+ "step": 820
616
+ },
617
+ {
618
+ "epoch": 1.04,
619
+ "grad_norm": 0.5248316526412964,
620
+ "learning_rate": 9.00513297924971e-06,
621
+ "loss": 0.5736,
622
+ "step": 830
623
+ },
624
+ {
625
+ "epoch": 1.05,
626
+ "grad_norm": 0.43092867732048035,
627
+ "learning_rate": 8.98129190120566e-06,
628
+ "loss": 0.5602,
629
+ "step": 840
630
+ },
631
+ {
632
+ "epoch": 1.07,
633
+ "grad_norm": 0.4282706379890442,
634
+ "learning_rate": 8.957200883538568e-06,
635
+ "loss": 0.5737,
636
+ "step": 850
637
+ },
638
+ {
639
+ "epoch": 1.08,
640
+ "grad_norm": 0.6257330775260925,
641
+ "learning_rate": 8.932861438646929e-06,
642
+ "loss": 0.5982,
643
+ "step": 860
644
+ },
645
+ {
646
+ "epoch": 1.09,
647
+ "grad_norm": 0.44131189584732056,
648
+ "learning_rate": 8.908275094525132e-06,
649
+ "loss": 0.6144,
650
+ "step": 870
651
+ },
652
+ {
653
+ "epoch": 1.1,
654
+ "grad_norm": 0.49996450543403625,
655
+ "learning_rate": 8.88344339466753e-06,
656
+ "loss": 0.6138,
657
+ "step": 880
658
+ },
659
+ {
660
+ "epoch": 1.12,
661
+ "grad_norm": 0.41754284501075745,
662
+ "learning_rate": 8.858367897971555e-06,
663
+ "loss": 0.5641,
664
+ "step": 890
665
+ },
666
+ {
667
+ "epoch": 1.13,
668
+ "grad_norm": 0.5091477036476135,
669
+ "learning_rate": 8.833050178639835e-06,
670
+ "loss": 0.5455,
671
+ "step": 900
672
+ },
673
+ {
674
+ "epoch": 1.14,
675
+ "grad_norm": 0.5106736421585083,
676
+ "learning_rate": 8.807491826081384e-06,
677
+ "loss": 0.5913,
678
+ "step": 910
679
+ },
680
+ {
681
+ "epoch": 1.15,
682
+ "grad_norm": 0.48033618927001953,
683
+ "learning_rate": 8.781694444811813e-06,
684
+ "loss": 0.5773,
685
+ "step": 920
686
+ },
687
+ {
688
+ "epoch": 1.17,
689
+ "grad_norm": 0.5638559460639954,
690
+ "learning_rate": 8.755659654352599e-06,
691
+ "loss": 0.5732,
692
+ "step": 930
693
+ },
694
+ {
695
+ "epoch": 1.18,
696
+ "grad_norm": 0.45605936646461487,
697
+ "learning_rate": 8.729389089129423e-06,
698
+ "loss": 0.5911,
699
+ "step": 940
700
+ },
701
+ {
702
+ "epoch": 1.19,
703
+ "grad_norm": 0.4803718328475952,
704
+ "learning_rate": 8.702884398369552e-06,
705
+ "loss": 0.5909,
706
+ "step": 950
707
+ },
708
+ {
709
+ "epoch": 1.2,
710
+ "grad_norm": 0.5215190649032593,
711
+ "learning_rate": 8.676147245998307e-06,
712
+ "loss": 0.614,
713
+ "step": 960
714
+ },
715
+ {
716
+ "epoch": 1.22,
717
+ "grad_norm": 0.490953654050827,
718
+ "learning_rate": 8.649179310534612e-06,
719
+ "loss": 0.5687,
720
+ "step": 970
721
+ },
722
+ {
723
+ "epoch": 1.23,
724
+ "grad_norm": 0.5034527778625488,
725
+ "learning_rate": 8.621982284985611e-06,
726
+ "loss": 0.5948,
727
+ "step": 980
728
+ },
729
+ {
730
+ "epoch": 1.24,
731
+ "grad_norm": 0.5091165900230408,
732
+ "learning_rate": 8.59455787674038e-06,
733
+ "loss": 0.5695,
734
+ "step": 990
735
+ },
736
+ {
737
+ "epoch": 1.25,
738
+ "grad_norm": 0.49061551690101624,
739
+ "learning_rate": 8.56690780746275e-06,
740
+ "loss": 0.5522,
741
+ "step": 1000
742
+ },
743
+ {
744
+ "epoch": 1.25,
745
+ "eval_loss": 0.5888673067092896,
746
+ "eval_runtime": 73.1303,
747
+ "eval_samples_per_second": 61.534,
748
+ "eval_steps_per_second": 1.928,
749
+ "step": 1000
750
+ },
751
+ {
752
+ "epoch": 1.27,
753
+ "grad_norm": 0.4834423065185547,
754
+ "learning_rate": 8.539033812983215e-06,
755
+ "loss": 0.5413,
756
+ "step": 1010
757
+ },
758
+ {
759
+ "epoch": 1.28,
760
+ "grad_norm": 0.6021168828010559,
761
+ "learning_rate": 8.510937643189962e-06,
762
+ "loss": 0.5882,
763
+ "step": 1020
764
+ },
765
+ {
766
+ "epoch": 1.29,
767
+ "grad_norm": 0.5302062630653381,
768
+ "learning_rate": 8.482621061919021e-06,
769
+ "loss": 0.6059,
770
+ "step": 1030
771
+ },
772
+ {
773
+ "epoch": 1.3,
774
+ "grad_norm": 0.4995095729827881,
775
+ "learning_rate": 8.454085846843523e-06,
776
+ "loss": 0.5419,
777
+ "step": 1040
778
+ },
779
+ {
780
+ "epoch": 1.32,
781
+ "grad_norm": 0.5463216304779053,
782
+ "learning_rate": 8.425333789362108e-06,
783
+ "loss": 0.5941,
784
+ "step": 1050
785
+ },
786
+ {
787
+ "epoch": 1.33,
788
+ "grad_norm": 0.536431610584259,
789
+ "learning_rate": 8.396366694486466e-06,
790
+ "loss": 0.6284,
791
+ "step": 1060
792
+ },
793
+ {
794
+ "epoch": 1.34,
795
+ "grad_norm": 0.5217477083206177,
796
+ "learning_rate": 8.36718638072801e-06,
797
+ "loss": 0.6178,
798
+ "step": 1070
799
+ },
800
+ {
801
+ "epoch": 1.36,
802
+ "grad_norm": 0.4792711138725281,
803
+ "learning_rate": 8.337794679983724e-06,
804
+ "loss": 0.5858,
805
+ "step": 1080
806
+ },
807
+ {
808
+ "epoch": 1.37,
809
+ "grad_norm": 0.5548479557037354,
810
+ "learning_rate": 8.30819343742115e-06,
811
+ "loss": 0.5861,
812
+ "step": 1090
813
+ },
814
+ {
815
+ "epoch": 1.38,
816
+ "grad_norm": 0.4717773199081421,
817
+ "learning_rate": 8.278384511362562e-06,
818
+ "loss": 0.6047,
819
+ "step": 1100
820
+ },
821
+ {
822
+ "epoch": 1.39,
823
+ "grad_norm": 0.5312036275863647,
824
+ "learning_rate": 8.248369773168297e-06,
825
+ "loss": 0.5994,
826
+ "step": 1110
827
+ },
828
+ {
829
+ "epoch": 1.41,
830
+ "grad_norm": 0.4545268714427948,
831
+ "learning_rate": 8.218151107119266e-06,
832
+ "loss": 0.5618,
833
+ "step": 1120
834
+ },
835
+ {
836
+ "epoch": 1.42,
837
+ "grad_norm": 0.6215177774429321,
838
+ "learning_rate": 8.187730410298675e-06,
839
+ "loss": 0.6228,
840
+ "step": 1130
841
+ },
842
+ {
843
+ "epoch": 1.43,
844
+ "grad_norm": 0.5464479923248291,
845
+ "learning_rate": 8.157109592472924e-06,
846
+ "loss": 0.5774,
847
+ "step": 1140
848
+ },
849
+ {
850
+ "epoch": 1.44,
851
+ "grad_norm": 0.5225650668144226,
852
+ "learning_rate": 8.126290575971714e-06,
853
+ "loss": 0.5804,
854
+ "step": 1150
855
+ },
856
+ {
857
+ "epoch": 1.46,
858
+ "grad_norm": 0.6188873648643494,
859
+ "learning_rate": 8.095275295567361e-06,
860
+ "loss": 0.5498,
861
+ "step": 1160
862
+ },
863
+ {
864
+ "epoch": 1.47,
865
+ "grad_norm": 0.5218258500099182,
866
+ "learning_rate": 8.064065698353344e-06,
867
+ "loss": 0.5829,
868
+ "step": 1170
869
+ },
870
+ {
871
+ "epoch": 1.48,
872
+ "grad_norm": 0.5987175703048706,
873
+ "learning_rate": 8.032663743622065e-06,
874
+ "loss": 0.5914,
875
+ "step": 1180
876
+ },
877
+ {
878
+ "epoch": 1.49,
879
+ "grad_norm": 0.527808666229248,
880
+ "learning_rate": 8.001071402741843e-06,
881
+ "loss": 0.5682,
882
+ "step": 1190
883
+ },
884
+ {
885
+ "epoch": 1.51,
886
+ "grad_norm": 0.5273358821868896,
887
+ "learning_rate": 7.969290659033156e-06,
888
+ "loss": 0.5919,
889
+ "step": 1200
890
+ },
891
+ {
892
+ "epoch": 1.51,
893
+ "eval_loss": 0.5836370587348938,
894
+ "eval_runtime": 71.4972,
895
+ "eval_samples_per_second": 62.939,
896
+ "eval_steps_per_second": 1.972,
897
+ "step": 1200
898
+ },
899
+ {
900
+ "epoch": 1.52,
901
+ "grad_norm": 0.6059889197349548,
902
+ "learning_rate": 7.93732350764414e-06,
903
+ "loss": 0.5763,
904
+ "step": 1210
905
+ },
906
+ {
907
+ "epoch": 1.53,
908
+ "grad_norm": 0.5306861996650696,
909
+ "learning_rate": 7.905171955425328e-06,
910
+ "loss": 0.5567,
911
+ "step": 1220
912
+ },
913
+ {
914
+ "epoch": 1.54,
915
+ "grad_norm": 0.5967408418655396,
916
+ "learning_rate": 7.872838020803662e-06,
917
+ "loss": 0.5609,
918
+ "step": 1230
919
+ },
920
+ {
921
+ "epoch": 1.56,
922
+ "grad_norm": 0.6726911067962646,
923
+ "learning_rate": 7.84032373365578e-06,
924
+ "loss": 0.6471,
925
+ "step": 1240
926
+ },
927
+ {
928
+ "epoch": 1.57,
929
+ "grad_norm": 0.5502099990844727,
930
+ "learning_rate": 7.807631135180591e-06,
931
+ "loss": 0.5886,
932
+ "step": 1250
933
+ },
934
+ {
935
+ "epoch": 1.58,
936
+ "grad_norm": 0.615433394908905,
937
+ "learning_rate": 7.774762277771123e-06,
938
+ "loss": 0.5559,
939
+ "step": 1260
940
+ },
941
+ {
942
+ "epoch": 1.59,
943
+ "grad_norm": 0.5481686592102051,
944
+ "learning_rate": 7.741719224885675e-06,
945
+ "loss": 0.5928,
946
+ "step": 1270
947
+ },
948
+ {
949
+ "epoch": 1.61,
950
+ "grad_norm": 0.620995044708252,
951
+ "learning_rate": 7.708504050918288e-06,
952
+ "loss": 0.5683,
953
+ "step": 1280
954
+ },
955
+ {
956
+ "epoch": 1.62,
957
+ "grad_norm": 0.5698446035385132,
958
+ "learning_rate": 7.675118841068507e-06,
959
+ "loss": 0.5909,
960
+ "step": 1290
961
+ },
962
+ {
963
+ "epoch": 1.63,
964
+ "grad_norm": 0.5968176126480103,
965
+ "learning_rate": 7.64156569121048e-06,
966
+ "loss": 0.573,
967
+ "step": 1300
968
+ },
969
+ {
970
+ "epoch": 1.64,
971
+ "grad_norm": 0.7233667373657227,
972
+ "learning_rate": 7.6078467077613746e-06,
973
+ "loss": 0.5888,
974
+ "step": 1310
975
+ },
976
+ {
977
+ "epoch": 1.66,
978
+ "grad_norm": 0.5293125510215759,
979
+ "learning_rate": 7.5739640075491546e-06,
980
+ "loss": 0.6316,
981
+ "step": 1320
982
+ },
983
+ {
984
+ "epoch": 1.67,
985
+ "grad_norm": 0.5775229334831238,
986
+ "learning_rate": 7.5399197176796735e-06,
987
+ "loss": 0.591,
988
+ "step": 1330
989
+ },
990
+ {
991
+ "epoch": 1.68,
992
+ "grad_norm": 0.7166782021522522,
993
+ "learning_rate": 7.50571597540315e-06,
994
+ "loss": 0.5505,
995
+ "step": 1340
996
+ },
997
+ {
998
+ "epoch": 1.69,
999
+ "grad_norm": 0.5287649631500244,
1000
+ "learning_rate": 7.471354927979984e-06,
1001
+ "loss": 0.6036,
1002
+ "step": 1350
1003
+ },
1004
+ {
1005
+ "epoch": 1.71,
1006
+ "grad_norm": 0.5856715440750122,
1007
+ "learning_rate": 7.436838732545964e-06,
1008
+ "loss": 0.5682,
1009
+ "step": 1360
1010
+ },
1011
+ {
1012
+ "epoch": 1.72,
1013
+ "grad_norm": 0.5226353406906128,
1014
+ "learning_rate": 7.402169555976841e-06,
1015
+ "loss": 0.5593,
1016
+ "step": 1370
1017
+ },
1018
+ {
1019
+ "epoch": 1.73,
1020
+ "grad_norm": 0.6625097393989563,
1021
+ "learning_rate": 7.367349574752296e-06,
1022
+ "loss": 0.5853,
1023
+ "step": 1380
1024
+ },
1025
+ {
1026
+ "epoch": 1.74,
1027
+ "grad_norm": 0.5979090929031372,
1028
+ "learning_rate": 7.332380974819301e-06,
1029
+ "loss": 0.5493,
1030
+ "step": 1390
1031
+ },
1032
+ {
1033
+ "epoch": 1.76,
1034
+ "grad_norm": 0.6176837682723999,
1035
+ "learning_rate": 7.297265951454891e-06,
1036
+ "loss": 0.6038,
1037
+ "step": 1400
1038
+ },
1039
+ {
1040
+ "epoch": 1.76,
1041
+ "eval_loss": 0.5800889730453491,
1042
+ "eval_runtime": 69.6671,
1043
+ "eval_samples_per_second": 64.593,
1044
+ "eval_steps_per_second": 2.024,
1045
+ "step": 1400
1046
+ },
1047
+ {
1048
+ "epoch": 1.77,
1049
+ "grad_norm": 0.6878402233123779,
1050
+ "learning_rate": 7.262006709128351e-06,
1051
+ "loss": 0.5453,
1052
+ "step": 1410
1053
+ },
1054
+ {
1055
+ "epoch": 1.78,
1056
+ "grad_norm": 0.6541485786437988,
1057
+ "learning_rate": 7.226605461362817e-06,
1058
+ "loss": 0.586,
1059
+ "step": 1420
1060
+ },
1061
+ {
1062
+ "epoch": 1.79,
1063
+ "grad_norm": 0.6143850088119507,
1064
+ "learning_rate": 7.191064430596315e-06,
1065
+ "loss": 0.5422,
1066
+ "step": 1430
1067
+ },
1068
+ {
1069
+ "epoch": 1.81,
1070
+ "grad_norm": 0.6676700115203857,
1071
+ "learning_rate": 7.155385848042245e-06,
1072
+ "loss": 0.5806,
1073
+ "step": 1440
1074
+ },
1075
+ {
1076
+ "epoch": 1.82,
1077
+ "grad_norm": 0.6332250833511353,
1078
+ "learning_rate": 7.119571953549305e-06,
1079
+ "loss": 0.5585,
1080
+ "step": 1450
1081
+ },
1082
+ {
1083
+ "epoch": 1.83,
1084
+ "grad_norm": 0.7023981809616089,
1085
+ "learning_rate": 7.08362499546087e-06,
1086
+ "loss": 0.5833,
1087
+ "step": 1460
1088
+ },
1089
+ {
1090
+ "epoch": 1.84,
1091
+ "grad_norm": 0.665985107421875,
1092
+ "learning_rate": 7.047547230473861e-06,
1093
+ "loss": 0.569,
1094
+ "step": 1470
1095
+ },
1096
+ {
1097
+ "epoch": 1.86,
1098
+ "grad_norm": 0.5501468777656555,
1099
+ "learning_rate": 7.011340923497055e-06,
1100
+ "loss": 0.5824,
1101
+ "step": 1480
1102
+ },
1103
+ {
1104
+ "epoch": 1.87,
1105
+ "grad_norm": 0.5913044810295105,
1106
+ "learning_rate": 6.975008347508906e-06,
1107
+ "loss": 0.5581,
1108
+ "step": 1490
1109
+ },
1110
+ {
1111
+ "epoch": 1.88,
1112
+ "grad_norm": 0.6394723653793335,
1113
+ "learning_rate": 6.938551783414856e-06,
1114
+ "loss": 0.5813,
1115
+ "step": 1500
1116
+ },
1117
+ {
1118
+ "epoch": 1.89,
1119
+ "grad_norm": 0.6261583566665649,
1120
+ "learning_rate": 6.901973519904128e-06,
1121
+ "loss": 0.5804,
1122
+ "step": 1510
1123
+ },
1124
+ {
1125
+ "epoch": 1.91,
1126
+ "grad_norm": 0.5661801695823669,
1127
+ "learning_rate": 6.865275853306062e-06,
1128
+ "loss": 0.5587,
1129
+ "step": 1520
1130
+ },
1131
+ {
1132
+ "epoch": 1.92,
1133
+ "grad_norm": 0.5941905975341797,
1134
+ "learning_rate": 6.828461087445946e-06,
1135
+ "loss": 0.5735,
1136
+ "step": 1530
1137
+ },
1138
+ {
1139
+ "epoch": 1.93,
1140
+ "grad_norm": 0.5802645087242126,
1141
+ "learning_rate": 6.791531533500384e-06,
1142
+ "loss": 0.5854,
1143
+ "step": 1540
1144
+ },
1145
+ {
1146
+ "epoch": 1.94,
1147
+ "grad_norm": 0.5715972781181335,
1148
+ "learning_rate": 6.754489509852208e-06,
1149
+ "loss": 0.5627,
1150
+ "step": 1550
1151
+ },
1152
+ {
1153
+ "epoch": 1.96,
1154
+ "grad_norm": 0.6669459342956543,
1155
+ "learning_rate": 6.7173373419449295e-06,
1156
+ "loss": 0.5647,
1157
+ "step": 1560
1158
+ },
1159
+ {
1160
+ "epoch": 1.97,
1161
+ "grad_norm": 0.6907923221588135,
1162
+ "learning_rate": 6.68007736213676e-06,
1163
+ "loss": 0.5364,
1164
+ "step": 1570
1165
+ },
1166
+ {
1167
+ "epoch": 1.98,
1168
+ "grad_norm": 0.5516065359115601,
1169
+ "learning_rate": 6.6427119095541745e-06,
1170
+ "loss": 0.5643,
1171
+ "step": 1580
1172
+ },
1173
+ {
1174
+ "epoch": 1.99,
1175
+ "grad_norm": 0.6715050935745239,
1176
+ "learning_rate": 6.60524332994508e-06,
1177
+ "loss": 0.6026,
1178
+ "step": 1590
1179
+ },
1180
+ {
1181
+ "epoch": 2.01,
1182
+ "grad_norm": 0.5935114026069641,
1183
+ "learning_rate": 6.567673975531542e-06,
1184
+ "loss": 0.5882,
1185
+ "step": 1600
1186
+ },
1187
+ {
1188
+ "epoch": 2.01,
1189
+ "eval_loss": 0.5773123502731323,
1190
+ "eval_runtime": 71.3304,
1191
+ "eval_samples_per_second": 63.087,
1192
+ "eval_steps_per_second": 1.977,
1193
+ "step": 1600
1194
+ },
1195
+ {
1196
+ "epoch": 2.02,
1197
+ "grad_norm": 0.7686386108398438,
1198
+ "learning_rate": 6.530006204862121e-06,
1199
+ "loss": 0.5391,
1200
+ "step": 1610
1201
+ },
1202
+ {
1203
+ "epoch": 2.03,
1204
+ "grad_norm": 0.6634293794631958,
1205
+ "learning_rate": 6.492242382663804e-06,
1206
+ "loss": 0.5495,
1207
+ "step": 1620
1208
+ },
1209
+ {
1210
+ "epoch": 2.05,
1211
+ "grad_norm": 0.8547865748405457,
1212
+ "learning_rate": 6.454384879693549e-06,
1213
+ "loss": 0.5456,
1214
+ "step": 1630
1215
+ },
1216
+ {
1217
+ "epoch": 2.06,
1218
+ "grad_norm": 0.8082049489021301,
1219
+ "learning_rate": 6.41643607258946e-06,
1220
+ "loss": 0.5903,
1221
+ "step": 1640
1222
+ },
1223
+ {
1224
+ "epoch": 2.07,
1225
+ "grad_norm": 0.6522872447967529,
1226
+ "learning_rate": 6.378398343721575e-06,
1227
+ "loss": 0.5476,
1228
+ "step": 1650
1229
+ },
1230
+ {
1231
+ "epoch": 2.08,
1232
+ "grad_norm": 0.663383960723877,
1233
+ "learning_rate": 6.340274081042307e-06,
1234
+ "loss": 0.5271,
1235
+ "step": 1660
1236
+ },
1237
+ {
1238
+ "epoch": 2.1,
1239
+ "grad_norm": 0.8022051453590393,
1240
+ "learning_rate": 6.3020656779365416e-06,
1241
+ "loss": 0.5656,
1242
+ "step": 1670
1243
+ },
1244
+ {
1245
+ "epoch": 2.11,
1246
+ "grad_norm": 0.7163932919502258,
1247
+ "learning_rate": 6.263775533071368e-06,
1248
+ "loss": 0.5581,
1249
+ "step": 1680
1250
+ },
1251
+ {
1252
+ "epoch": 2.12,
1253
+ "grad_norm": 0.6052596569061279,
1254
+ "learning_rate": 6.225406050245507e-06,
1255
+ "loss": 0.554,
1256
+ "step": 1690
1257
+ },
1258
+ {
1259
+ "epoch": 2.13,
1260
+ "grad_norm": 0.6838263273239136,
1261
+ "learning_rate": 6.186959638238397e-06,
1262
+ "loss": 0.5812,
1263
+ "step": 1700
1264
+ },
1265
+ {
1266
+ "epoch": 2.15,
1267
+ "grad_norm": 0.6921390891075134,
1268
+ "learning_rate": 6.148438710658979e-06,
1269
+ "loss": 0.5237,
1270
+ "step": 1710
1271
+ },
1272
+ {
1273
+ "epoch": 2.16,
1274
+ "grad_norm": 0.883646547794342,
1275
+ "learning_rate": 6.109845685794169e-06,
1276
+ "loss": 0.5267,
1277
+ "step": 1720
1278
+ },
1279
+ {
1280
+ "epoch": 2.17,
1281
+ "grad_norm": 0.6446242928504944,
1282
+ "learning_rate": 6.071182986457046e-06,
1283
+ "loss": 0.5466,
1284
+ "step": 1730
1285
+ },
1286
+ {
1287
+ "epoch": 2.18,
1288
+ "grad_norm": 0.7913056015968323,
1289
+ "learning_rate": 6.032453039834748e-06,
1290
+ "loss": 0.573,
1291
+ "step": 1740
1292
+ },
1293
+ {
1294
+ "epoch": 2.2,
1295
+ "grad_norm": 0.7966864705085754,
1296
+ "learning_rate": 5.993658277336101e-06,
1297
+ "loss": 0.5498,
1298
+ "step": 1750
1299
+ },
1300
+ {
1301
+ "epoch": 2.21,
1302
+ "grad_norm": 0.7100770473480225,
1303
+ "learning_rate": 5.954801134438971e-06,
1304
+ "loss": 0.5624,
1305
+ "step": 1760
1306
+ },
1307
+ {
1308
+ "epoch": 2.22,
1309
+ "grad_norm": 0.5544886589050293,
1310
+ "learning_rate": 5.915884050537378e-06,
1311
+ "loss": 0.5349,
1312
+ "step": 1770
1313
+ },
1314
+ {
1315
+ "epoch": 2.23,
1316
+ "grad_norm": 0.7371806502342224,
1317
+ "learning_rate": 5.8769094687883484e-06,
1318
+ "loss": 0.5789,
1319
+ "step": 1780
1320
+ },
1321
+ {
1322
+ "epoch": 2.25,
1323
+ "grad_norm": 0.7266082763671875,
1324
+ "learning_rate": 5.837879835958535e-06,
1325
+ "loss": 0.5573,
1326
+ "step": 1790
1327
+ },
1328
+ {
1329
+ "epoch": 2.26,
1330
+ "grad_norm": 0.6665763854980469,
1331
+ "learning_rate": 5.798797602270623e-06,
1332
+ "loss": 0.5302,
1333
+ "step": 1800
1334
+ },
1335
+ {
1336
+ "epoch": 2.26,
1337
+ "eval_loss": 0.5754047632217407,
1338
+ "eval_runtime": 72.9642,
1339
+ "eval_samples_per_second": 61.674,
1340
+ "eval_steps_per_second": 1.932,
1341
+ "step": 1800
1342
+ },
1343
+ {
1344
+ "epoch": 2.27,
1345
+ "grad_norm": 0.7089454531669617,
1346
+ "learning_rate": 5.7596652212494944e-06,
1347
+ "loss": 0.5652,
1348
+ "step": 1810
1349
+ },
1350
+ {
1351
+ "epoch": 2.28,
1352
+ "grad_norm": 0.790558397769928,
1353
+ "learning_rate": 5.720485149568211e-06,
1354
+ "loss": 0.5547,
1355
+ "step": 1820
1356
+ },
1357
+ {
1358
+ "epoch": 2.3,
1359
+ "grad_norm": 0.7344884872436523,
1360
+ "learning_rate": 5.681259846893783e-06,
1361
+ "loss": 0.5703,
1362
+ "step": 1830
1363
+ },
1364
+ {
1365
+ "epoch": 2.31,
1366
+ "grad_norm": 0.7799204587936401,
1367
+ "learning_rate": 5.641991775732756e-06,
1368
+ "loss": 0.5369,
1369
+ "step": 1840
1370
+ },
1371
+ {
1372
+ "epoch": 2.32,
1373
+ "grad_norm": 0.7437064051628113,
1374
+ "learning_rate": 5.6026834012766155e-06,
1375
+ "loss": 0.5814,
1376
+ "step": 1850
1377
+ },
1378
+ {
1379
+ "epoch": 2.33,
1380
+ "grad_norm": 0.7207692861557007,
1381
+ "learning_rate": 5.5633371912470334e-06,
1382
+ "loss": 0.5535,
1383
+ "step": 1860
1384
+ },
1385
+ {
1386
+ "epoch": 2.35,
1387
+ "grad_norm": 0.7650673985481262,
1388
+ "learning_rate": 5.5239556157409366e-06,
1389
+ "loss": 0.5507,
1390
+ "step": 1870
1391
+ },
1392
+ {
1393
+ "epoch": 2.36,
1394
+ "grad_norm": 0.8299598097801208,
1395
+ "learning_rate": 5.484541147075449e-06,
1396
+ "loss": 0.5678,
1397
+ "step": 1880
1398
+ },
1399
+ {
1400
+ "epoch": 2.37,
1401
+ "grad_norm": 0.6018086671829224,
1402
+ "learning_rate": 5.4450962596326776e-06,
1403
+ "loss": 0.5749,
1404
+ "step": 1890
1405
+ },
1406
+ {
1407
+ "epoch": 2.38,
1408
+ "grad_norm": 0.7759122848510742,
1409
+ "learning_rate": 5.405623429704373e-06,
1410
+ "loss": 0.5561,
1411
+ "step": 1900
1412
+ },
1413
+ {
1414
+ "epoch": 2.4,
1415
+ "grad_norm": 0.7874650359153748,
1416
+ "learning_rate": 5.366125135336477e-06,
1417
+ "loss": 0.568,
1418
+ "step": 1910
1419
+ },
1420
+ {
1421
+ "epoch": 2.41,
1422
+ "grad_norm": 0.642218828201294,
1423
+ "learning_rate": 5.326603856173548e-06,
1424
+ "loss": 0.5468,
1425
+ "step": 1920
1426
+ },
1427
+ {
1428
+ "epoch": 2.42,
1429
+ "grad_norm": 0.7558509707450867,
1430
+ "learning_rate": 5.2870620733030994e-06,
1431
+ "loss": 0.5315,
1432
+ "step": 1930
1433
+ },
1434
+ {
1435
+ "epoch": 2.43,
1436
+ "grad_norm": 0.846817672252655,
1437
+ "learning_rate": 5.247502269099834e-06,
1438
+ "loss": 0.5511,
1439
+ "step": 1940
1440
+ },
1441
+ {
1442
+ "epoch": 2.45,
1443
+ "grad_norm": 0.725319504737854,
1444
+ "learning_rate": 5.207926927069809e-06,
1445
+ "loss": 0.5098,
1446
+ "step": 1950
1447
+ },
1448
+ {
1449
+ "epoch": 2.46,
1450
+ "grad_norm": 0.7795179486274719,
1451
+ "learning_rate": 5.168338531694523e-06,
1452
+ "loss": 0.5417,
1453
+ "step": 1960
1454
+ },
1455
+ {
1456
+ "epoch": 2.47,
1457
+ "grad_norm": 0.764240562915802,
1458
+ "learning_rate": 5.1287395682749444e-06,
1459
+ "loss": 0.5502,
1460
+ "step": 1970
1461
+ },
1462
+ {
1463
+ "epoch": 2.48,
1464
+ "grad_norm": 0.7871197462081909,
1465
+ "learning_rate": 5.089132522775486e-06,
1466
+ "loss": 0.575,
1467
+ "step": 1980
1468
+ },
1469
+ {
1470
+ "epoch": 2.5,
1471
+ "grad_norm": 0.8167646527290344,
1472
+ "learning_rate": 5.049519881667945e-06,
1473
+ "loss": 0.5637,
1474
+ "step": 1990
1475
+ },
1476
+ {
1477
+ "epoch": 2.51,
1478
+ "grad_norm": 0.7696388959884644,
1479
+ "learning_rate": 5.009904131775402e-06,
1480
+ "loss": 0.5331,
1481
+ "step": 2000
1482
+ },
1483
+ {
1484
+ "epoch": 2.51,
1485
+ "eval_loss": 0.5732148289680481,
1486
+ "eval_runtime": 73.9926,
1487
+ "eval_samples_per_second": 60.817,
1488
+ "eval_steps_per_second": 1.906,
1489
+ "step": 2000
1490
+ },
1491
+ {
1492
+ "epoch": 2.52,
1493
+ "grad_norm": 0.8611487746238708,
1494
+ "learning_rate": 4.970287760116098e-06,
1495
+ "loss": 0.544,
1496
+ "step": 2010
1497
+ },
1498
+ {
1499
+ "epoch": 2.53,
1500
+ "grad_norm": 0.7960777878761292,
1501
+ "learning_rate": 4.930673253747314e-06,
1502
+ "loss": 0.5702,
1503
+ "step": 2020
1504
+ },
1505
+ {
1506
+ "epoch": 2.55,
1507
+ "grad_norm": 0.8783027529716492,
1508
+ "learning_rate": 4.8910630996092265e-06,
1509
+ "loss": 0.5309,
1510
+ "step": 2030
1511
+ },
1512
+ {
1513
+ "epoch": 2.56,
1514
+ "grad_norm": 0.7091485261917114,
1515
+ "learning_rate": 4.851459784368787e-06,
1516
+ "loss": 0.5565,
1517
+ "step": 2040
1518
+ },
1519
+ {
1520
+ "epoch": 2.57,
1521
+ "grad_norm": 0.7412302494049072,
1522
+ "learning_rate": 4.811865794263611e-06,
1523
+ "loss": 0.5546,
1524
+ "step": 2050
1525
+ },
1526
+ {
1527
+ "epoch": 2.58,
1528
+ "grad_norm": 0.7188848257064819,
1529
+ "learning_rate": 4.772283614945895e-06,
1530
+ "loss": 0.5927,
1531
+ "step": 2060
1532
+ },
1533
+ {
1534
+ "epoch": 2.6,
1535
+ "grad_norm": 0.9045892953872681,
1536
+ "learning_rate": 4.732715731326372e-06,
1537
+ "loss": 0.5515,
1538
+ "step": 2070
1539
+ },
1540
+ {
1541
+ "epoch": 2.61,
1542
+ "grad_norm": 0.7781686186790466,
1543
+ "learning_rate": 4.693164627418315e-06,
1544
+ "loss": 0.5289,
1545
+ "step": 2080
1546
+ },
1547
+ {
1548
+ "epoch": 2.62,
1549
+ "grad_norm": 0.8868610858917236,
1550
+ "learning_rate": 4.653632786181587e-06,
1551
+ "loss": 0.5934,
1552
+ "step": 2090
1553
+ },
1554
+ {
1555
+ "epoch": 2.63,
1556
+ "grad_norm": 0.7511059641838074,
1557
+ "learning_rate": 4.614122689366769e-06,
1558
+ "loss": 0.5464,
1559
+ "step": 2100
1560
+ },
1561
+ {
1562
+ "epoch": 2.65,
1563
+ "grad_norm": 0.7453913688659668,
1564
+ "learning_rate": 4.574636817359365e-06,
1565
+ "loss": 0.5318,
1566
+ "step": 2110
1567
+ },
1568
+ {
1569
+ "epoch": 2.66,
1570
+ "grad_norm": 0.6865512132644653,
1571
+ "learning_rate": 4.535177649024077e-06,
1572
+ "loss": 0.5527,
1573
+ "step": 2120
1574
+ },
1575
+ {
1576
+ "epoch": 2.67,
1577
+ "grad_norm": 0.8288848400115967,
1578
+ "learning_rate": 4.495747661549187e-06,
1579
+ "loss": 0.5368,
1580
+ "step": 2130
1581
+ },
1582
+ {
1583
+ "epoch": 2.69,
1584
+ "grad_norm": 0.8635351657867432,
1585
+ "learning_rate": 4.456349330291055e-06,
1586
+ "loss": 0.5549,
1587
+ "step": 2140
1588
+ },
1589
+ {
1590
+ "epoch": 2.7,
1591
+ "grad_norm": 0.7543601393699646,
1592
+ "learning_rate": 4.4169851286187005e-06,
1593
+ "loss": 0.5435,
1594
+ "step": 2150
1595
+ },
1596
+ {
1597
+ "epoch": 2.71,
1598
+ "grad_norm": 0.8399900794029236,
1599
+ "learning_rate": 4.3776575277585445e-06,
1600
+ "loss": 0.5853,
1601
+ "step": 2160
1602
+ },
1603
+ {
1604
+ "epoch": 2.72,
1605
+ "grad_norm": 0.9056334495544434,
1606
+ "learning_rate": 4.3383689966392604e-06,
1607
+ "loss": 0.5737,
1608
+ "step": 2170
1609
+ },
1610
+ {
1611
+ "epoch": 2.74,
1612
+ "grad_norm": 0.7865429520606995,
1613
+ "learning_rate": 4.2991220017367795e-06,
1614
+ "loss": 0.5908,
1615
+ "step": 2180
1616
+ },
1617
+ {
1618
+ "epoch": 2.75,
1619
+ "grad_norm": 0.9085346460342407,
1620
+ "learning_rate": 4.259919006919456e-06,
1621
+ "loss": 0.5468,
1622
+ "step": 2190
1623
+ },
1624
+ {
1625
+ "epoch": 2.76,
1626
+ "grad_norm": 0.844716489315033,
1627
+ "learning_rate": 4.220762473293379e-06,
1628
+ "loss": 0.5669,
1629
+ "step": 2200
1630
+ },
1631
+ {
1632
+ "epoch": 2.76,
1633
+ "eval_loss": 0.5714393854141235,
1634
+ "eval_runtime": 74.9812,
1635
+ "eval_samples_per_second": 60.015,
1636
+ "eval_steps_per_second": 1.88,
1637
+ "step": 2200
1638
+ },
1639
+ {
1640
+ "epoch": 2.77,
1641
+ "grad_norm": 0.8216796517372131,
1642
+ "learning_rate": 4.181654859047873e-06,
1643
+ "loss": 0.5516,
1644
+ "step": 2210
1645
+ },
1646
+ {
1647
+ "epoch": 2.79,
1648
+ "grad_norm": 0.8482044339179993,
1649
+ "learning_rate": 4.1425986193011805e-06,
1650
+ "loss": 0.5463,
1651
+ "step": 2220
1652
+ },
1653
+ {
1654
+ "epoch": 2.8,
1655
+ "grad_norm": 0.8037722706794739,
1656
+ "learning_rate": 4.103596205946323e-06,
1657
+ "loss": 0.5387,
1658
+ "step": 2230
1659
+ },
1660
+ {
1661
+ "epoch": 2.81,
1662
+ "grad_norm": 0.8523209691047668,
1663
+ "learning_rate": 4.0646500674971826e-06,
1664
+ "loss": 0.5282,
1665
+ "step": 2240
1666
+ },
1667
+ {
1668
+ "epoch": 2.82,
1669
+ "grad_norm": 0.8726175427436829,
1670
+ "learning_rate": 4.0257626489347895e-06,
1671
+ "loss": 0.5476,
1672
+ "step": 2250
1673
+ },
1674
+ {
1675
+ "epoch": 2.84,
1676
+ "grad_norm": 0.9041292071342468,
1677
+ "learning_rate": 3.986936391553822e-06,
1678
+ "loss": 0.5677,
1679
+ "step": 2260
1680
+ },
1681
+ {
1682
+ "epoch": 2.85,
1683
+ "grad_norm": 0.8939766883850098,
1684
+ "learning_rate": 3.948173732809351e-06,
1685
+ "loss": 0.5711,
1686
+ "step": 2270
1687
+ },
1688
+ {
1689
+ "epoch": 2.86,
1690
+ "grad_norm": 0.7777726650238037,
1691
+ "learning_rate": 3.909477106163818e-06,
1692
+ "loss": 0.5289,
1693
+ "step": 2280
1694
+ },
1695
+ {
1696
+ "epoch": 2.87,
1697
+ "grad_norm": 0.7699301242828369,
1698
+ "learning_rate": 3.87084894093427e-06,
1699
+ "loss": 0.55,
1700
+ "step": 2290
1701
+ },
1702
+ {
1703
+ "epoch": 2.89,
1704
+ "grad_norm": 0.9905028939247131,
1705
+ "learning_rate": 3.832291662139845e-06,
1706
+ "loss": 0.541,
1707
+ "step": 2300
1708
+ },
1709
+ {
1710
+ "epoch": 2.9,
1711
+ "grad_norm": 0.8651571869850159,
1712
+ "learning_rate": 3.7938076903495363e-06,
1713
+ "loss": 0.5533,
1714
+ "step": 2310
1715
+ },
1716
+ {
1717
+ "epoch": 2.91,
1718
+ "grad_norm": 0.7639217376708984,
1719
+ "learning_rate": 3.7553994415302274e-06,
1720
+ "loss": 0.5511,
1721
+ "step": 2320
1722
+ },
1723
+ {
1724
+ "epoch": 2.92,
1725
+ "grad_norm": 0.8610684871673584,
1726
+ "learning_rate": 3.7170693268950354e-06,
1727
+ "loss": 0.5119,
1728
+ "step": 2330
1729
+ },
1730
+ {
1731
+ "epoch": 2.94,
1732
+ "grad_norm": 0.7909624576568604,
1733
+ "learning_rate": 3.67881975275192e-06,
1734
+ "loss": 0.5881,
1735
+ "step": 2340
1736
+ },
1737
+ {
1738
+ "epoch": 2.95,
1739
+ "grad_norm": 0.9253331422805786,
1740
+ "learning_rate": 3.6406531203526274e-06,
1741
+ "loss": 0.5612,
1742
+ "step": 2350
1743
+ },
1744
+ {
1745
+ "epoch": 2.96,
1746
+ "grad_norm": 0.8581608533859253,
1747
+ "learning_rate": 3.6025718257419532e-06,
1748
+ "loss": 0.5322,
1749
+ "step": 2360
1750
+ },
1751
+ {
1752
+ "epoch": 2.97,
1753
+ "grad_norm": 0.9016821384429932,
1754
+ "learning_rate": 3.564578259607303e-06,
1755
+ "loss": 0.5487,
1756
+ "step": 2370
1757
+ },
1758
+ {
1759
+ "epoch": 2.99,
1760
+ "grad_norm": 0.8990395665168762,
1761
+ "learning_rate": 3.5266748071286205e-06,
1762
+ "loss": 0.5471,
1763
+ "step": 2380
1764
+ },
1765
+ {
1766
+ "epoch": 3.0,
1767
+ "grad_norm": 0.9548435211181641,
1768
+ "learning_rate": 3.488863847828652e-06,
1769
+ "loss": 0.547,
1770
+ "step": 2390
1771
+ },
1772
+ {
1773
+ "epoch": 3.01,
1774
+ "grad_norm": 0.771557092666626,
1775
+ "learning_rate": 3.4511477554235516e-06,
1776
+ "loss": 0.5551,
1777
+ "step": 2400
1778
+ },
1779
+ {
1780
+ "epoch": 3.01,
1781
+ "eval_loss": 0.5705960392951965,
1782
+ "eval_runtime": 76.5829,
1783
+ "eval_samples_per_second": 58.76,
1784
+ "eval_steps_per_second": 1.841,
1785
+ "step": 2400
1786
+ },
1787
+ {
1788
+ "epoch": 3.02,
1789
+ "grad_norm": 0.8474746942520142,
1790
+ "learning_rate": 3.4135288976738762e-06,
1791
+ "loss": 0.5552,
1792
+ "step": 2410
1793
+ },
1794
+ {
1795
+ "epoch": 3.04,
1796
+ "grad_norm": 1.0370606184005737,
1797
+ "learning_rate": 3.3760096362359306e-06,
1798
+ "loss": 0.5456,
1799
+ "step": 2420
1800
+ },
1801
+ {
1802
+ "epoch": 3.05,
1803
+ "grad_norm": 0.7640136480331421,
1804
+ "learning_rate": 3.33859232651351e-06,
1805
+ "loss": 0.5228,
1806
+ "step": 2430
1807
+ },
1808
+ {
1809
+ "epoch": 3.06,
1810
+ "grad_norm": 1.0315712690353394,
1811
+ "learning_rate": 3.301279317510038e-06,
1812
+ "loss": 0.5389,
1813
+ "step": 2440
1814
+ },
1815
+ {
1816
+ "epoch": 3.07,
1817
+ "grad_norm": 1.0031884908676147,
1818
+ "learning_rate": 3.2640729516810844e-06,
1819
+ "loss": 0.5153,
1820
+ "step": 2450
1821
+ },
1822
+ {
1823
+ "epoch": 3.09,
1824
+ "grad_norm": 0.8636874556541443,
1825
+ "learning_rate": 3.226975564787322e-06,
1826
+ "loss": 0.527,
1827
+ "step": 2460
1828
+ },
1829
+ {
1830
+ "epoch": 3.1,
1831
+ "grad_norm": 0.9102957248687744,
1832
+ "learning_rate": 3.1899894857478897e-06,
1833
+ "loss": 0.5397,
1834
+ "step": 2470
1835
+ },
1836
+ {
1837
+ "epoch": 3.11,
1838
+ "grad_norm": 0.8529887795448303,
1839
+ "learning_rate": 3.15311703649418e-06,
1840
+ "loss": 0.5184,
1841
+ "step": 2480
1842
+ },
1843
+ {
1844
+ "epoch": 3.12,
1845
+ "grad_norm": 0.9462995529174805,
1846
+ "learning_rate": 3.116360531824074e-06,
1847
+ "loss": 0.5559,
1848
+ "step": 2490
1849
+ },
1850
+ {
1851
+ "epoch": 3.14,
1852
+ "grad_norm": 0.8420120477676392,
1853
+ "learning_rate": 3.0797222792566306e-06,
1854
+ "loss": 0.5245,
1855
+ "step": 2500
1856
+ },
1857
+ {
1858
+ "epoch": 3.15,
1859
+ "grad_norm": 0.7965397238731384,
1860
+ "learning_rate": 3.043204578887209e-06,
1861
+ "loss": 0.5585,
1862
+ "step": 2510
1863
+ },
1864
+ {
1865
+ "epoch": 3.16,
1866
+ "grad_norm": 0.9633888602256775,
1867
+ "learning_rate": 3.00680972324308e-06,
1868
+ "loss": 0.4956,
1869
+ "step": 2520
1870
+ },
1871
+ {
1872
+ "epoch": 3.17,
1873
+ "grad_norm": 1.1136802434921265,
1874
+ "learning_rate": 2.970539997139509e-06,
1875
+ "loss": 0.5444,
1876
+ "step": 2530
1877
+ },
1878
+ {
1879
+ "epoch": 3.19,
1880
+ "grad_norm": 0.8797339797019958,
1881
+ "learning_rate": 2.9343976775363097e-06,
1882
+ "loss": 0.5253,
1883
+ "step": 2540
1884
+ },
1885
+ {
1886
+ "epoch": 3.2,
1887
+ "grad_norm": 1.0063321590423584,
1888
+ "learning_rate": 2.8983850333949065e-06,
1889
+ "loss": 0.5427,
1890
+ "step": 2550
1891
+ },
1892
+ {
1893
+ "epoch": 3.21,
1894
+ "grad_norm": 0.8839291930198669,
1895
+ "learning_rate": 2.862504325535886e-06,
1896
+ "loss": 0.5346,
1897
+ "step": 2560
1898
+ },
1899
+ {
1900
+ "epoch": 3.22,
1901
+ "grad_norm": 0.9732418060302734,
1902
+ "learning_rate": 2.8267578064970735e-06,
1903
+ "loss": 0.5889,
1904
+ "step": 2570
1905
+ },
1906
+ {
1907
+ "epoch": 3.24,
1908
+ "grad_norm": 1.0936757326126099,
1909
+ "learning_rate": 2.791147720392118e-06,
1910
+ "loss": 0.5136,
1911
+ "step": 2580
1912
+ },
1913
+ {
1914
+ "epoch": 3.25,
1915
+ "grad_norm": 0.9694124460220337,
1916
+ "learning_rate": 2.755676302769611e-06,
1917
+ "loss": 0.5276,
1918
+ "step": 2590
1919
+ },
1920
+ {
1921
+ "epoch": 3.26,
1922
+ "grad_norm": 0.9088903069496155,
1923
+ "learning_rate": 2.7203457804727364e-06,
1924
+ "loss": 0.5499,
1925
+ "step": 2600
1926
+ },
1927
+ {
1928
+ "epoch": 3.26,
1929
+ "eval_loss": 0.5715650320053101,
1930
+ "eval_runtime": 75.4356,
1931
+ "eval_samples_per_second": 59.653,
1932
+ "eval_steps_per_second": 1.869,
1933
+ "step": 2600
1934
+ },
1935
+ {
1936
+ "epoch": 3.27,
1937
+ "grad_norm": 0.8832049369812012,
1938
+ "learning_rate": 2.6851583714994843e-06,
1939
+ "loss": 0.5392,
1940
+ "step": 2610
1941
+ },
1942
+ {
1943
+ "epoch": 3.29,
1944
+ "grad_norm": 1.0563007593154907,
1945
+ "learning_rate": 2.6501162848634023e-06,
1946
+ "loss": 0.5952,
1947
+ "step": 2620
1948
+ },
1949
+ {
1950
+ "epoch": 3.3,
1951
+ "grad_norm": 0.9844443798065186,
1952
+ "learning_rate": 2.6152217204549045e-06,
1953
+ "loss": 0.5304,
1954
+ "step": 2630
1955
+ },
1956
+ {
1957
+ "epoch": 3.31,
1958
+ "grad_norm": 1.0739080905914307,
1959
+ "learning_rate": 2.580476868903197e-06,
1960
+ "loss": 0.4941,
1961
+ "step": 2640
1962
+ },
1963
+ {
1964
+ "epoch": 3.32,
1965
+ "grad_norm": 0.8901020884513855,
1966
+ "learning_rate": 2.5458839114387146e-06,
1967
+ "loss": 0.5418,
1968
+ "step": 2650
1969
+ },
1970
+ {
1971
+ "epoch": 3.34,
1972
+ "grad_norm": 0.9322318434715271,
1973
+ "learning_rate": 2.5114450197562214e-06,
1974
+ "loss": 0.4952,
1975
+ "step": 2660
1976
+ },
1977
+ {
1978
+ "epoch": 3.35,
1979
+ "grad_norm": 0.9946108460426331,
1980
+ "learning_rate": 2.4771623558784474e-06,
1981
+ "loss": 0.5155,
1982
+ "step": 2670
1983
+ },
1984
+ {
1985
+ "epoch": 3.36,
1986
+ "grad_norm": 0.9886655211448669,
1987
+ "learning_rate": 2.4430380720203785e-06,
1988
+ "loss": 0.5374,
1989
+ "step": 2680
1990
+ },
1991
+ {
1992
+ "epoch": 3.38,
1993
+ "grad_norm": 0.9973739385604858,
1994
+ "learning_rate": 2.409074310454136e-06,
1995
+ "loss": 0.5055,
1996
+ "step": 2690
1997
+ },
1998
+ {
1999
+ "epoch": 3.39,
2000
+ "grad_norm": 0.9865736961364746,
2001
+ "learning_rate": 2.375273203374489e-06,
2002
+ "loss": 0.5287,
2003
+ "step": 2700
2004
+ },
2005
+ {
2006
+ "epoch": 3.4,
2007
+ "grad_norm": 0.9992399215698242,
2008
+ "learning_rate": 2.341636872764994e-06,
2009
+ "loss": 0.5753,
2010
+ "step": 2710
2011
+ },
2012
+ {
2013
+ "epoch": 3.41,
2014
+ "grad_norm": 1.0160070657730103,
2015
+ "learning_rate": 2.3081674302647905e-06,
2016
+ "loss": 0.5102,
2017
+ "step": 2720
2018
+ },
2019
+ {
2020
+ "epoch": 3.43,
2021
+ "grad_norm": 0.7992890477180481,
2022
+ "learning_rate": 2.2748669770360273e-06,
2023
+ "loss": 0.5551,
2024
+ "step": 2730
2025
+ },
2026
+ {
2027
+ "epoch": 3.44,
2028
+ "grad_norm": 1.0403869152069092,
2029
+ "learning_rate": 2.2417376036319495e-06,
2030
+ "loss": 0.5231,
2031
+ "step": 2740
2032
+ },
2033
+ {
2034
+ "epoch": 3.45,
2035
+ "grad_norm": 0.9026238918304443,
2036
+ "learning_rate": 2.2087813898656775e-06,
2037
+ "loss": 0.5148,
2038
+ "step": 2750
2039
+ },
2040
+ {
2041
+ "epoch": 3.46,
2042
+ "grad_norm": 1.078641414642334,
2043
+ "learning_rate": 2.176000404679612e-06,
2044
+ "loss": 0.5481,
2045
+ "step": 2760
2046
+ },
2047
+ {
2048
+ "epoch": 3.48,
2049
+ "grad_norm": 0.8496494889259338,
2050
+ "learning_rate": 2.1433967060155687e-06,
2051
+ "loss": 0.4961,
2052
+ "step": 2770
2053
+ },
2054
+ {
2055
+ "epoch": 3.49,
2056
+ "grad_norm": 0.9248107671737671,
2057
+ "learning_rate": 2.110972340685577e-06,
2058
+ "loss": 0.5402,
2059
+ "step": 2780
2060
+ },
2061
+ {
2062
+ "epoch": 3.5,
2063
+ "grad_norm": 1.0704095363616943,
2064
+ "learning_rate": 2.078729344243384e-06,
2065
+ "loss": 0.5228,
2066
+ "step": 2790
2067
+ },
2068
+ {
2069
+ "epoch": 3.51,
2070
+ "grad_norm": 0.8805019855499268,
2071
+ "learning_rate": 2.0466697408566664e-06,
2072
+ "loss": 0.545,
2073
+ "step": 2800
2074
+ },
2075
+ {
2076
+ "epoch": 3.51,
2077
+ "eval_loss": 0.5704147815704346,
2078
+ "eval_runtime": 77.346,
2079
+ "eval_samples_per_second": 58.18,
2080
+ "eval_steps_per_second": 1.823,
2081
+ "step": 2800
2082
+ },
2083
+ {
2084
+ "epoch": 3.53,
2085
+ "grad_norm": 1.0108119249343872,
2086
+ "learning_rate": 2.014795543179957e-06,
2087
+ "loss": 0.5361,
2088
+ "step": 2810
2089
+ },
2090
+ {
2091
+ "epoch": 3.54,
2092
+ "grad_norm": 1.01576828956604,
2093
+ "learning_rate": 1.9831087522282865e-06,
2094
+ "loss": 0.5552,
2095
+ "step": 2820
2096
+ },
2097
+ {
2098
+ "epoch": 3.55,
2099
+ "grad_norm": 0.9220107197761536,
2100
+ "learning_rate": 1.951611357251575e-06,
2101
+ "loss": 0.5753,
2102
+ "step": 2830
2103
+ },
2104
+ {
2105
+ "epoch": 3.56,
2106
+ "grad_norm": 0.9482864141464233,
2107
+ "learning_rate": 1.92030533560974e-06,
2108
+ "loss": 0.5339,
2109
+ "step": 2840
2110
+ },
2111
+ {
2112
+ "epoch": 3.58,
2113
+ "grad_norm": 0.9235343933105469,
2114
+ "learning_rate": 1.8891926526485643e-06,
2115
+ "loss": 0.538,
2116
+ "step": 2850
2117
+ },
2118
+ {
2119
+ "epoch": 3.59,
2120
+ "grad_norm": 0.6921684145927429,
2121
+ "learning_rate": 1.8582752615763161e-06,
2122
+ "loss": 0.5331,
2123
+ "step": 2860
2124
+ },
2125
+ {
2126
+ "epoch": 3.6,
2127
+ "grad_norm": 0.930234968662262,
2128
+ "learning_rate": 1.8275551033411209e-06,
2129
+ "loss": 0.5226,
2130
+ "step": 2870
2131
+ },
2132
+ {
2133
+ "epoch": 3.61,
2134
+ "grad_norm": 0.9894780516624451,
2135
+ "learning_rate": 1.7970341065091246e-06,
2136
+ "loss": 0.5294,
2137
+ "step": 2880
2138
+ },
2139
+ {
2140
+ "epoch": 3.63,
2141
+ "grad_norm": 1.0227577686309814,
2142
+ "learning_rate": 1.766714187143413e-06,
2143
+ "loss": 0.5186,
2144
+ "step": 2890
2145
+ },
2146
+ {
2147
+ "epoch": 3.64,
2148
+ "grad_norm": 0.9115012884140015,
2149
+ "learning_rate": 1.736597248683724e-06,
2150
+ "loss": 0.5156,
2151
+ "step": 2900
2152
+ },
2153
+ {
2154
+ "epoch": 3.65,
2155
+ "grad_norm": 1.050728440284729,
2156
+ "learning_rate": 1.7066851818269558e-06,
2157
+ "loss": 0.521,
2158
+ "step": 2910
2159
+ },
2160
+ {
2161
+ "epoch": 3.66,
2162
+ "grad_norm": 1.0954750776290894,
2163
+ "learning_rate": 1.676979864408471e-06,
2164
+ "loss": 0.5451,
2165
+ "step": 2920
2166
+ },
2167
+ {
2168
+ "epoch": 3.68,
2169
+ "grad_norm": 0.7569468021392822,
2170
+ "learning_rate": 1.647483161284203e-06,
2171
+ "loss": 0.5765,
2172
+ "step": 2930
2173
+ },
2174
+ {
2175
+ "epoch": 3.69,
2176
+ "grad_norm": 0.8553792238235474,
2177
+ "learning_rate": 1.6181969242135926e-06,
2178
+ "loss": 0.5335,
2179
+ "step": 2940
2180
+ },
2181
+ {
2182
+ "epoch": 3.7,
2183
+ "grad_norm": 0.918138861656189,
2184
+ "learning_rate": 1.5891229917433316e-06,
2185
+ "loss": 0.5343,
2186
+ "step": 2950
2187
+ },
2188
+ {
2189
+ "epoch": 3.71,
2190
+ "grad_norm": 0.988727867603302,
2191
+ "learning_rate": 1.5602631890919428e-06,
2192
+ "loss": 0.5288,
2193
+ "step": 2960
2194
+ },
2195
+ {
2196
+ "epoch": 3.73,
2197
+ "grad_norm": 1.130867600440979,
2198
+ "learning_rate": 1.531619328035197e-06,
2199
+ "loss": 0.4952,
2200
+ "step": 2970
2201
+ },
2202
+ {
2203
+ "epoch": 3.74,
2204
+ "grad_norm": 1.1033731698989868,
2205
+ "learning_rate": 1.503193206792366e-06,
2206
+ "loss": 0.5053,
2207
+ "step": 2980
2208
+ },
2209
+ {
2210
+ "epoch": 3.75,
2211
+ "grad_norm": 1.0361392498016357,
2212
+ "learning_rate": 1.4749866099133409e-06,
2213
+ "loss": 0.555,
2214
+ "step": 2990
2215
+ },
2216
+ {
2217
+ "epoch": 3.76,
2218
+ "grad_norm": 1.1811940670013428,
2219
+ "learning_rate": 1.447001308166599e-06,
2220
+ "loss": 0.5253,
2221
+ "step": 3000
2222
+ },
2223
+ {
2224
+ "epoch": 3.76,
2225
+ "eval_loss": 0.5704381465911865,
2226
+ "eval_runtime": 72.5701,
2227
+ "eval_samples_per_second": 62.009,
2228
+ "eval_steps_per_second": 1.943,
2229
+ "step": 3000
2230
+ },
2231
+ {
2232
+ "epoch": 3.78,
2233
+ "grad_norm": 1.0649921894073486,
2234
+ "learning_rate": 1.4192390584280347e-06,
2235
+ "loss": 0.5397,
2236
+ "step": 3010
2237
+ },
2238
+ {
2239
+ "epoch": 3.79,
2240
+ "grad_norm": 1.006032943725586,
2241
+ "learning_rate": 1.3917016035706632e-06,
2242
+ "loss": 0.5321,
2243
+ "step": 3020
2244
+ },
2245
+ {
2246
+ "epoch": 3.8,
2247
+ "grad_norm": 1.1129940748214722,
2248
+ "learning_rate": 1.3643906723552186e-06,
2249
+ "loss": 0.5246,
2250
+ "step": 3030
2251
+ },
2252
+ {
2253
+ "epoch": 3.81,
2254
+ "grad_norm": 1.1166002750396729,
2255
+ "learning_rate": 1.3373079793216037e-06,
2256
+ "loss": 0.5405,
2257
+ "step": 3040
2258
+ },
2259
+ {
2260
+ "epoch": 3.83,
2261
+ "grad_norm": 0.9260038137435913,
2262
+ "learning_rate": 1.3104552246812768e-06,
2263
+ "loss": 0.543,
2264
+ "step": 3050
2265
+ },
2266
+ {
2267
+ "epoch": 3.84,
2268
+ "grad_norm": 1.0943797826766968,
2269
+ "learning_rate": 1.2838340942104944e-06,
2270
+ "loss": 0.5147,
2271
+ "step": 3060
2272
+ },
2273
+ {
2274
+ "epoch": 3.85,
2275
+ "grad_norm": 1.0058125257492065,
2276
+ "learning_rate": 1.257446259144494e-06,
2277
+ "loss": 0.4995,
2278
+ "step": 3070
2279
+ },
2280
+ {
2281
+ "epoch": 3.86,
2282
+ "grad_norm": 1.022620677947998,
2283
+ "learning_rate": 1.2312933760725753e-06,
2284
+ "loss": 0.5485,
2285
+ "step": 3080
2286
+ },
2287
+ {
2288
+ "epoch": 3.88,
2289
+ "grad_norm": 1.077143907546997,
2290
+ "learning_rate": 1.2053770868340914e-06,
2291
+ "loss": 0.5483,
2292
+ "step": 3090
2293
+ },
2294
+ {
2295
+ "epoch": 3.89,
2296
+ "grad_norm": 1.3038913011550903,
2297
+ "learning_rate": 1.17969901841539e-06,
2298
+ "loss": 0.5388,
2299
+ "step": 3100
2300
+ },
2301
+ {
2302
+ "epoch": 3.9,
2303
+ "grad_norm": 0.8922839760780334,
2304
+ "learning_rate": 1.1542607828476665e-06,
2305
+ "loss": 0.5404,
2306
+ "step": 3110
2307
+ },
2308
+ {
2309
+ "epoch": 3.91,
2310
+ "grad_norm": 0.9709892272949219,
2311
+ "learning_rate": 1.1290639771057648e-06,
2312
+ "loss": 0.5204,
2313
+ "step": 3120
2314
+ },
2315
+ {
2316
+ "epoch": 3.93,
2317
+ "grad_norm": 1.1010524034500122,
2318
+ "learning_rate": 1.1041101830079159e-06,
2319
+ "loss": 0.5429,
2320
+ "step": 3130
2321
+ },
2322
+ {
2323
+ "epoch": 3.94,
2324
+ "grad_norm": 1.1041542291641235,
2325
+ "learning_rate": 1.0794009671164484e-06,
2326
+ "loss": 0.5256,
2327
+ "step": 3140
2328
+ },
2329
+ {
2330
+ "epoch": 3.95,
2331
+ "grad_norm": 1.1114507913589478,
2332
+ "learning_rate": 1.054937880639424e-06,
2333
+ "loss": 0.5185,
2334
+ "step": 3150
2335
+ },
2336
+ {
2337
+ "epoch": 3.96,
2338
+ "grad_norm": 0.9926356077194214,
2339
+ "learning_rate": 1.030722459333267e-06,
2340
+ "loss": 0.5161,
2341
+ "step": 3160
2342
+ },
2343
+ {
2344
+ "epoch": 3.98,
2345
+ "grad_norm": 1.00425386428833,
2346
+ "learning_rate": 1.0067562234063477e-06,
2347
+ "loss": 0.5222,
2348
+ "step": 3170
2349
+ },
2350
+ {
2351
+ "epoch": 3.99,
2352
+ "grad_norm": 1.0224225521087646,
2353
+ "learning_rate": 9.830406774235474e-07,
2354
+ "loss": 0.521,
2355
+ "step": 3180
2356
+ },
2357
+ {
2358
+ "epoch": 4.0,
2359
+ "grad_norm": 0.9822995662689209,
2360
+ "learning_rate": 9.595773102118022e-07,
2361
+ "loss": 0.528,
2362
+ "step": 3190
2363
+ },
2364
+ {
2365
+ "epoch": 4.02,
2366
+ "grad_norm": 1.1153786182403564,
2367
+ "learning_rate": 9.363675947666362e-07,
2368
+ "loss": 0.5418,
2369
+ "step": 3200
2370
+ },
2371
+ {
2372
+ "epoch": 4.02,
2373
+ "eval_loss": 0.5702469944953918,
2374
+ "eval_runtime": 71.9125,
2375
+ "eval_samples_per_second": 62.576,
2376
+ "eval_steps_per_second": 1.961,
2377
+ "step": 3200
2378
+ },
2379
+ {
2380
+ "epoch": 4.03,
2381
+ "grad_norm": 1.056578278541565,
2382
+ "learning_rate": 9.134129881596926e-07,
2383
+ "loss": 0.5326,
2384
+ "step": 3210
2385
+ },
2386
+ {
2387
+ "epoch": 4.04,
2388
+ "grad_norm": 0.9958868026733398,
2389
+ "learning_rate": 8.907149314472579e-07,
2390
+ "loss": 0.558,
2391
+ "step": 3220
2392
+ },
2393
+ {
2394
+ "epoch": 4.05,
2395
+ "grad_norm": 1.1032071113586426,
2396
+ "learning_rate": 8.682748495797965e-07,
2397
+ "loss": 0.5423,
2398
+ "step": 3230
2399
+ },
2400
+ {
2401
+ "epoch": 4.07,
2402
+ "grad_norm": 1.0517561435699463,
2403
+ "learning_rate": 8.460941513124876e-07,
2404
+ "loss": 0.5189,
2405
+ "step": 3240
2406
+ },
2407
+ {
2408
+ "epoch": 4.08,
2409
+ "grad_norm": 1.1289842128753662,
2410
+ "learning_rate": 8.241742291168015e-07,
2411
+ "loss": 0.5002,
2412
+ "step": 3250
2413
+ },
2414
+ {
2415
+ "epoch": 4.09,
2416
+ "grad_norm": 0.9873390793800354,
2417
+ "learning_rate": 8.025164590930623e-07,
2418
+ "loss": 0.5589,
2419
+ "step": 3260
2420
+ },
2421
+ {
2422
+ "epoch": 4.1,
2423
+ "grad_norm": 0.885833740234375,
2424
+ "learning_rate": 7.811222008840719e-07,
2425
+ "loss": 0.5031,
2426
+ "step": 3270
2427
+ },
2428
+ {
2429
+ "epoch": 4.12,
2430
+ "grad_norm": 0.7554776072502136,
2431
+ "learning_rate": 7.599927975897514e-07,
2432
+ "loss": 0.5128,
2433
+ "step": 3280
2434
+ },
2435
+ {
2436
+ "epoch": 4.13,
2437
+ "grad_norm": 1.0165417194366455,
2438
+ "learning_rate": 7.391295756828187e-07,
2439
+ "loss": 0.5369,
2440
+ "step": 3290
2441
+ },
2442
+ {
2443
+ "epoch": 4.14,
2444
+ "grad_norm": 0.9958201050758362,
2445
+ "learning_rate": 7.185338449255197e-07,
2446
+ "loss": 0.4948,
2447
+ "step": 3300
2448
+ },
2449
+ {
2450
+ "epoch": 4.15,
2451
+ "grad_norm": 1.1157667636871338,
2452
+ "learning_rate": 6.982068982874019e-07,
2453
+ "loss": 0.5181,
2454
+ "step": 3310
2455
+ },
2456
+ {
2457
+ "epoch": 4.17,
2458
+ "grad_norm": 1.0274876356124878,
2459
+ "learning_rate": 6.781500118641377e-07,
2460
+ "loss": 0.5172,
2461
+ "step": 3320
2462
+ },
2463
+ {
2464
+ "epoch": 4.18,
2465
+ "grad_norm": 0.9985108375549316,
2466
+ "learning_rate": 6.583644447974241e-07,
2467
+ "loss": 0.5304,
2468
+ "step": 3330
2469
+ },
2470
+ {
2471
+ "epoch": 4.19,
2472
+ "grad_norm": 1.0142953395843506,
2473
+ "learning_rate": 6.388514391959283e-07,
2474
+ "loss": 0.5278,
2475
+ "step": 3340
2476
+ },
2477
+ {
2478
+ "epoch": 4.2,
2479
+ "grad_norm": 1.0171020030975342,
2480
+ "learning_rate": 6.196122200573073e-07,
2481
+ "loss": 0.5269,
2482
+ "step": 3350
2483
+ },
2484
+ {
2485
+ "epoch": 4.22,
2486
+ "grad_norm": 1.2027987241744995,
2487
+ "learning_rate": 6.006479951913168e-07,
2488
+ "loss": 0.5379,
2489
+ "step": 3360
2490
+ },
2491
+ {
2492
+ "epoch": 4.23,
2493
+ "grad_norm": 1.064353108406067,
2494
+ "learning_rate": 5.819599551439703e-07,
2495
+ "loss": 0.4879,
2496
+ "step": 3370
2497
+ },
2498
+ {
2499
+ "epoch": 4.24,
2500
+ "grad_norm": 1.0336158275604248,
2501
+ "learning_rate": 5.635492731228098e-07,
2502
+ "loss": 0.5363,
2503
+ "step": 3380
2504
+ },
2505
+ {
2506
+ "epoch": 4.25,
2507
+ "grad_norm": 1.143526554107666,
2508
+ "learning_rate": 5.45417104923251e-07,
2509
+ "loss": 0.5371,
2510
+ "step": 3390
2511
+ },
2512
+ {
2513
+ "epoch": 4.27,
2514
+ "grad_norm": 0.9199160933494568,
2515
+ "learning_rate": 5.275645888560233e-07,
2516
+ "loss": 0.5289,
2517
+ "step": 3400
2518
+ },
2519
+ {
2520
+ "epoch": 4.27,
2521
+ "eval_loss": 0.570408046245575,
2522
+ "eval_runtime": 72.3442,
2523
+ "eval_samples_per_second": 62.203,
2524
+ "eval_steps_per_second": 1.949,
2525
+ "step": 3400
2526
+ },
2527
+ {
2528
+ "epoch": 4.28,
2529
+ "grad_norm": 1.0219011306762695,
2530
+ "learning_rate": 5.099928456757036e-07,
2531
+ "loss": 0.5115,
2532
+ "step": 3410
2533
+ },
2534
+ {
2535
+ "epoch": 4.29,
2536
+ "grad_norm": 1.0187703371047974,
2537
+ "learning_rate": 4.927029785103715e-07,
2538
+ "loss": 0.5171,
2539
+ "step": 3420
2540
+ },
2541
+ {
2542
+ "epoch": 4.3,
2543
+ "grad_norm": 1.006913423538208,
2544
+ "learning_rate": 4.756960727923371e-07,
2545
+ "loss": 0.5113,
2546
+ "step": 3430
2547
+ },
2548
+ {
2549
+ "epoch": 4.32,
2550
+ "grad_norm": 1.0249481201171875,
2551
+ "learning_rate": 4.5897319619001404e-07,
2552
+ "loss": 0.552,
2553
+ "step": 3440
2554
+ },
2555
+ {
2556
+ "epoch": 4.33,
2557
+ "grad_norm": 1.016176462173462,
2558
+ "learning_rate": 4.42535398540887e-07,
2559
+ "loss": 0.5246,
2560
+ "step": 3450
2561
+ },
2562
+ {
2563
+ "epoch": 4.34,
2564
+ "grad_norm": 0.9975274205207825,
2565
+ "learning_rate": 4.263837117856046e-07,
2566
+ "loss": 0.5328,
2567
+ "step": 3460
2568
+ },
2569
+ {
2570
+ "epoch": 4.35,
2571
+ "grad_norm": 0.8603295683860779,
2572
+ "learning_rate": 4.10519149903198e-07,
2573
+ "loss": 0.5409,
2574
+ "step": 3470
2575
+ },
2576
+ {
2577
+ "epoch": 4.37,
2578
+ "grad_norm": 1.079770565032959,
2579
+ "learning_rate": 3.949427088474189e-07,
2580
+ "loss": 0.5506,
2581
+ "step": 3480
2582
+ },
2583
+ {
2584
+ "epoch": 4.38,
2585
+ "grad_norm": 0.9214140176773071,
2586
+ "learning_rate": 3.79655366484224e-07,
2587
+ "loss": 0.5175,
2588
+ "step": 3490
2589
+ },
2590
+ {
2591
+ "epoch": 4.39,
2592
+ "grad_norm": 0.9992761015892029,
2593
+ "learning_rate": 3.646580825303786e-07,
2594
+ "loss": 0.5197,
2595
+ "step": 3500
2596
+ },
2597
+ {
2598
+ "epoch": 4.4,
2599
+ "grad_norm": 0.9221498370170593,
2600
+ "learning_rate": 3.49951798493211e-07,
2601
+ "loss": 0.5083,
2602
+ "step": 3510
2603
+ },
2604
+ {
2605
+ "epoch": 4.42,
2606
+ "grad_norm": 1.0544548034667969,
2607
+ "learning_rate": 3.355374376115017e-07,
2608
+ "loss": 0.5187,
2609
+ "step": 3520
2610
+ },
2611
+ {
2612
+ "epoch": 4.43,
2613
+ "grad_norm": 0.8989390134811401,
2614
+ "learning_rate": 3.214159047975324e-07,
2615
+ "loss": 0.5137,
2616
+ "step": 3530
2617
+ },
2618
+ {
2619
+ "epoch": 4.44,
2620
+ "grad_norm": 0.9407683610916138,
2621
+ "learning_rate": 3.0758808658026496e-07,
2622
+ "loss": 0.5045,
2623
+ "step": 3540
2624
+ },
2625
+ {
2626
+ "epoch": 4.45,
2627
+ "grad_norm": 1.034885287284851,
2628
+ "learning_rate": 2.940548510496977e-07,
2629
+ "loss": 0.5369,
2630
+ "step": 3550
2631
+ },
2632
+ {
2633
+ "epoch": 4.47,
2634
+ "grad_norm": 0.9151519536972046,
2635
+ "learning_rate": 2.8081704780236083e-07,
2636
+ "loss": 0.5619,
2637
+ "step": 3560
2638
+ },
2639
+ {
2640
+ "epoch": 4.48,
2641
+ "grad_norm": 1.22360098361969,
2642
+ "learning_rate": 2.6787550788798243e-07,
2643
+ "loss": 0.5078,
2644
+ "step": 3570
2645
+ },
2646
+ {
2647
+ "epoch": 4.49,
2648
+ "grad_norm": 1.11081063747406,
2649
+ "learning_rate": 2.55231043757318e-07,
2650
+ "loss": 0.5391,
2651
+ "step": 3580
2652
+ },
2653
+ {
2654
+ "epoch": 4.5,
2655
+ "grad_norm": 0.9431936144828796,
2656
+ "learning_rate": 2.428844492111415e-07,
2657
+ "loss": 0.581,
2658
+ "step": 3590
2659
+ },
2660
+ {
2661
+ "epoch": 4.52,
2662
+ "grad_norm": 0.993818461894989,
2663
+ "learning_rate": 2.3083649935041642e-07,
2664
+ "loss": 0.4983,
2665
+ "step": 3600
2666
+ },
2667
+ {
2668
+ "epoch": 4.52,
2669
+ "eval_loss": 0.570551335811615,
2670
+ "eval_runtime": 73.4513,
2671
+ "eval_samples_per_second": 61.265,
2672
+ "eval_steps_per_second": 1.92,
2673
+ "step": 3600
2674
+ },
2675
+ {
2676
+ "epoch": 4.53,
2677
+ "grad_norm": 0.9568440318107605,
2678
+ "learning_rate": 2.1908795052763388e-07,
2679
+ "loss": 0.5094,
2680
+ "step": 3610
2681
+ },
2682
+ {
2683
+ "epoch": 4.54,
2684
+ "grad_norm": 1.0978604555130005,
2685
+ "learning_rate": 2.0763954029933085e-07,
2686
+ "loss": 0.5262,
2687
+ "step": 3620
2688
+ },
2689
+ {
2690
+ "epoch": 4.55,
2691
+ "grad_norm": 0.9831346273422241,
2692
+ "learning_rate": 1.9649198737978325e-07,
2693
+ "loss": 0.5381,
2694
+ "step": 3630
2695
+ },
2696
+ {
2697
+ "epoch": 4.57,
2698
+ "grad_norm": 1.1549886465072632,
2699
+ "learning_rate": 1.856459915958958e-07,
2700
+ "loss": 0.5069,
2701
+ "step": 3640
2702
+ },
2703
+ {
2704
+ "epoch": 4.58,
2705
+ "grad_norm": 1.122920274734497,
2706
+ "learning_rate": 1.7510223384325674e-07,
2707
+ "loss": 0.5323,
2708
+ "step": 3650
2709
+ },
2710
+ {
2711
+ "epoch": 4.59,
2712
+ "grad_norm": 1.0226640701293945,
2713
+ "learning_rate": 1.6486137604339813e-07,
2714
+ "loss": 0.5206,
2715
+ "step": 3660
2716
+ },
2717
+ {
2718
+ "epoch": 4.6,
2719
+ "grad_norm": 0.9887501001358032,
2720
+ "learning_rate": 1.5492406110224289e-07,
2721
+ "loss": 0.549,
2722
+ "step": 3670
2723
+ },
2724
+ {
2725
+ "epoch": 4.62,
2726
+ "grad_norm": 1.069175124168396,
2727
+ "learning_rate": 1.4529091286973994e-07,
2728
+ "loss": 0.535,
2729
+ "step": 3680
2730
+ },
2731
+ {
2732
+ "epoch": 4.63,
2733
+ "grad_norm": 1.0605347156524658,
2734
+ "learning_rate": 1.3596253610070165e-07,
2735
+ "loss": 0.4874,
2736
+ "step": 3690
2737
+ },
2738
+ {
2739
+ "epoch": 4.64,
2740
+ "grad_norm": 1.1193597316741943,
2741
+ "learning_rate": 1.2693951641683856e-07,
2742
+ "loss": 0.5137,
2743
+ "step": 3700
2744
+ },
2745
+ {
2746
+ "epoch": 4.65,
2747
+ "grad_norm": 1.0010859966278076,
2748
+ "learning_rate": 1.1822242026999443e-07,
2749
+ "loss": 0.4834,
2750
+ "step": 3710
2751
+ },
2752
+ {
2753
+ "epoch": 4.67,
2754
+ "grad_norm": 1.1622557640075684,
2755
+ "learning_rate": 1.0981179490658522e-07,
2756
+ "loss": 0.5739,
2757
+ "step": 3720
2758
+ },
2759
+ {
2760
+ "epoch": 4.68,
2761
+ "grad_norm": 1.0148837566375732,
2762
+ "learning_rate": 1.0170816833324549e-07,
2763
+ "loss": 0.5182,
2764
+ "step": 3730
2765
+ },
2766
+ {
2767
+ "epoch": 4.69,
2768
+ "grad_norm": 1.1132534742355347,
2769
+ "learning_rate": 9.391204928367648e-08,
2770
+ "loss": 0.517,
2771
+ "step": 3740
2772
+ },
2773
+ {
2774
+ "epoch": 4.71,
2775
+ "grad_norm": 1.0367366075515747,
2776
+ "learning_rate": 8.642392718671455e-08,
2777
+ "loss": 0.5191,
2778
+ "step": 3750
2779
+ },
2780
+ {
2781
+ "epoch": 4.72,
2782
+ "grad_norm": 1.0188932418823242,
2783
+ "learning_rate": 7.92442721355996e-08,
2784
+ "loss": 0.5277,
2785
+ "step": 3760
2786
+ },
2787
+ {
2788
+ "epoch": 4.73,
2789
+ "grad_norm": 1.0666061639785767,
2790
+ "learning_rate": 7.237353485846865e-08,
2791
+ "loss": 0.5272,
2792
+ "step": 3770
2793
+ },
2794
+ {
2795
+ "epoch": 4.74,
2796
+ "grad_norm": 0.9185099601745605,
2797
+ "learning_rate": 6.58121466900552e-08,
2798
+ "loss": 0.5114,
2799
+ "step": 3780
2800
+ },
2801
+ {
2802
+ "epoch": 4.76,
2803
+ "grad_norm": 1.1004579067230225,
2804
+ "learning_rate": 5.9560519544614725e-08,
2805
+ "loss": 0.5088,
2806
+ "step": 3790
2807
+ },
2808
+ {
2809
+ "epoch": 4.77,
2810
+ "grad_norm": 1.0939706563949585,
2811
+ "learning_rate": 5.36190458900615e-08,
2812
+ "loss": 0.5371,
2813
+ "step": 3800
2814
+ },
2815
+ {
2816
+ "epoch": 4.77,
2817
+ "eval_loss": 0.5705362558364868,
2818
+ "eval_runtime": 72.0164,
2819
+ "eval_samples_per_second": 62.486,
2820
+ "eval_steps_per_second": 1.958,
2821
+ "step": 3800
2822
+ },
2823
+ {
2824
+ "epoch": 4.78,
2825
+ "grad_norm": 1.0084924697875977,
2826
+ "learning_rate": 4.798809872333276e-08,
2827
+ "loss": 0.5561,
2828
+ "step": 3810
2829
+ },
2830
+ {
2831
+ "epoch": 4.79,
2832
+ "grad_norm": 1.118712067604065,
2833
+ "learning_rate": 4.266803154697241e-08,
2834
+ "loss": 0.5511,
2835
+ "step": 3820
2836
+ },
2837
+ {
2838
+ "epoch": 4.81,
2839
+ "grad_norm": 1.0696793794631958,
2840
+ "learning_rate": 3.765917834693711e-08,
2841
+ "loss": 0.4753,
2842
+ "step": 3830
2843
+ },
2844
+ {
2845
+ "epoch": 4.82,
2846
+ "grad_norm": 0.9693195223808289,
2847
+ "learning_rate": 3.296185357162973e-08,
2848
+ "loss": 0.5308,
2849
+ "step": 3840
2850
+ },
2851
+ {
2852
+ "epoch": 4.83,
2853
+ "grad_norm": 1.145647406578064,
2854
+ "learning_rate": 2.8576352112158478e-08,
2855
+ "loss": 0.5174,
2856
+ "step": 3850
2857
+ },
2858
+ {
2859
+ "epoch": 4.84,
2860
+ "grad_norm": 0.9660029411315918,
2861
+ "learning_rate": 2.4502949283826683e-08,
2862
+ "loss": 0.4866,
2863
+ "step": 3860
2864
+ },
2865
+ {
2866
+ "epoch": 4.86,
2867
+ "grad_norm": 0.9987422823905945,
2868
+ "learning_rate": 2.074190080884386e-08,
2869
+ "loss": 0.5169,
2870
+ "step": 3870
2871
+ },
2872
+ {
2873
+ "epoch": 4.87,
2874
+ "grad_norm": 0.9800373911857605,
2875
+ "learning_rate": 1.729344280027745e-08,
2876
+ "loss": 0.5065,
2877
+ "step": 3880
2878
+ },
2879
+ {
2880
+ "epoch": 4.88,
2881
+ "grad_norm": 1.1359976530075073,
2882
+ "learning_rate": 1.4157791747225202e-08,
2883
+ "loss": 0.5496,
2884
+ "step": 3890
2885
+ },
2886
+ {
2887
+ "epoch": 4.89,
2888
+ "grad_norm": 0.9586722254753113,
2889
+ "learning_rate": 1.133514450122719e-08,
2890
+ "loss": 0.5138,
2891
+ "step": 3900
2892
+ },
2893
+ {
2894
+ "epoch": 4.91,
2895
+ "grad_norm": 1.1620326042175293,
2896
+ "learning_rate": 8.825678263906789e-09,
2897
+ "loss": 0.5573,
2898
+ "step": 3910
2899
+ },
2900
+ {
2901
+ "epoch": 4.92,
2902
+ "grad_norm": 0.8677152395248413,
2903
+ "learning_rate": 6.629550575847355e-09,
2904
+ "loss": 0.4891,
2905
+ "step": 3920
2906
+ },
2907
+ {
2908
+ "epoch": 4.93,
2909
+ "grad_norm": 1.0054067373275757,
2910
+ "learning_rate": 4.746899306699582e-09,
2911
+ "loss": 0.5325,
2912
+ "step": 3930
2913
+ },
2914
+ {
2915
+ "epoch": 4.94,
2916
+ "grad_norm": 1.2144650220870972,
2917
+ "learning_rate": 3.1778426465289838e-09,
2918
+ "loss": 0.4983,
2919
+ "step": 3940
2920
+ },
2921
+ {
2922
+ "epoch": 4.96,
2923
+ "grad_norm": 1.1614487171173096,
2924
+ "learning_rate": 1.9224790983940477e-09,
2925
+ "loss": 0.5108,
2926
+ "step": 3950
2927
+ },
2928
+ {
2929
+ "epoch": 4.97,
2930
+ "grad_norm": 0.9379814267158508,
2931
+ "learning_rate": 9.808874721628503e-10,
2932
+ "loss": 0.5122,
2933
+ "step": 3960
2934
+ },
2935
+ {
2936
+ "epoch": 4.98,
2937
+ "grad_norm": 1.0148200988769531,
2938
+ "learning_rate": 3.5312687956756684e-10,
2939
+ "loss": 0.5337,
2940
+ "step": 3970
2941
+ },
2942
+ {
2943
+ "epoch": 4.99,
2944
+ "grad_norm": 1.089341640472412,
2945
+ "learning_rate": 3.923673048966681e-11,
2946
+ "loss": 0.529,
2947
+ "step": 3980
2948
+ },
2949
+ {
2950
+ "epoch": 5.0,
2951
+ "step": 3985,
2952
+ "total_flos": 2.2824671149817856e+18,
2953
+ "train_loss": 0.569244503227052,
2954
+ "train_runtime": 8741.9073,
2955
+ "train_samples_per_second": 14.585,
2956
+ "train_steps_per_second": 0.456
2957
+ }
2958
+ ],
2959
+ "logging_steps": 10,
2960
+ "max_steps": 3985,
2961
+ "num_input_tokens_seen": 0,
2962
+ "num_train_epochs": 5,
2963
+ "save_steps": 1000,
2964
+ "total_flos": 2.2824671149817856e+18,
2965
+ "train_batch_size": 16,
2966
+ "trial_name": null,
2967
+ "trial_params": null
2968
+ }
llama2_13b_peft/news_commentary_es/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f945155f44888fffb9611998b5a2cb53a0ecf7e194d1ec8ea3f4941fc7e5b93
3
+ size 5112
llama2_13b_peft/news_commentary_es/training_eval_loss.png ADDED
llama2_13b_peft/news_commentary_es/training_loss.png ADDED
llama2_13b_peft/object_counting/README.md ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ library_name: peft
4
+ tags:
5
+ - llama-factory
6
+ - lora
7
+ - generated_from_trainer
8
+ base_model: /data1/model/llama2/meta-llama/Llama2-13b
9
+ model-index:
10
+ - name: object_counting_no_sys
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # object_counting_no_sys
18
+
19
+ This model is a fine-tuned version of [/data1/model/llama2/meta-llama/Llama2-13b](https://huggingface.co//data1/model/llama2/meta-llama/Llama2-13b) on the object_counting_no_sys dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.2107
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0001
41
+ - train_batch_size: 4
42
+ - eval_batch_size: 4
43
+ - seed: 42
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: cosine
46
+ - lr_scheduler_warmup_steps: 20
47
+ - num_epochs: 5.0
48
+
49
+ ### Training results
50
+
51
+ | Training Loss | Epoch | Step | Validation Loss |
52
+ |:-------------:|:------:|:----:|:---------------:|
53
+ | 0.477 | 0.5263 | 100 | 0.4622 |
54
+ | 0.4647 | 1.0526 | 200 | 0.2513 |
55
+ | 0.5231 | 1.5789 | 300 | 0.3432 |
56
+ | 0.4381 | 2.1053 | 400 | 0.2107 |
57
+ | 0.2159 | 2.6316 | 500 | 0.3666 |
58
+ | 0.1934 | 3.1579 | 600 | 0.2625 |
59
+ | 0.2701 | 3.6842 | 700 | 0.2854 |
60
+
61
+
62
+ ### Framework versions
63
+
64
+ - PEFT 0.10.0
65
+ - Transformers 4.40.0
66
+ - Pytorch 2.2.1
67
+ - Datasets 2.18.0
68
+ - Tokenizers 0.19.1
llama2_13b_peft/object_counting/adapter_config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data1/model/llama2/meta-llama/Llama2-13b",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 16,
14
+ "lora_dropout": 0.0,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "gate_proj",
24
+ "v_proj",
25
+ "q_proj",
26
+ "up_proj",
27
+ "o_proj",
28
+ "k_proj",
29
+ "down_proj"
30
+ ],
31
+ "task_type": "CAUSAL_LM",
32
+ "use_dora": false,
33
+ "use_rslora": false
34
+ }
llama2_13b_peft/object_counting/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ec85f92f9090e17b7da8fd52d43bbdc1cfa184a3c8bd57e86f55941c22c564d
3
+ size 125248064
llama2_13b_peft/object_counting/all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.6842105263157894,
3
+ "eval_loss": 0.21073441207408905,
4
+ "eval_runtime": 0.6962,
5
+ "eval_samples_per_second": 57.458,
6
+ "eval_steps_per_second": 14.365,
7
+ "total_flos": 1.461153753759744e+16,
8
+ "train_loss": 0.5859067852156503,
9
+ "train_runtime": 192.1858,
10
+ "train_samples_per_second": 19.773,
11
+ "train_steps_per_second": 4.943
12
+ }
llama2_13b_peft/object_counting/eval_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.6842105263157894,
3
+ "eval_loss": 0.21073441207408905,
4
+ "eval_runtime": 0.6962,
5
+ "eval_samples_per_second": 57.458,
6
+ "eval_steps_per_second": 14.365
7
+ }
llama2_13b_peft/object_counting/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
llama2_13b_peft/object_counting/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
llama2_13b_peft/object_counting/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content + '\\n' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }
llama2_13b_peft/object_counting/train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.6842105263157894,
3
+ "total_flos": 1.461153753759744e+16,
4
+ "train_loss": 0.5859067852156503,
5
+ "train_runtime": 192.1858,
6
+ "train_samples_per_second": 19.773,
7
+ "train_steps_per_second": 4.943
8
+ }
llama2_13b_peft/object_counting/trainer_log.jsonl ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"current_steps": 10, "total_steps": 950, "loss": 10.645, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.05263157894736842, "percentage": 1.05, "elapsed_time": "0:00:03", "remaining_time": "0:06:06"}
2
+ {"current_steps": 20, "total_steps": 950, "loss": 4.9002, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 0.0001, "epoch": 0.10526315789473684, "percentage": 2.11, "elapsed_time": "0:00:06", "remaining_time": "0:04:58"}
3
+ {"current_steps": 30, "total_steps": 950, "loss": 0.9012, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.99714745464859e-05, "epoch": 0.15789473684210525, "percentage": 3.16, "elapsed_time": "0:00:09", "remaining_time": "0:04:36"}
4
+ {"current_steps": 40, "total_steps": 950, "loss": 0.7022, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.988593073400354e-05, "epoch": 0.21052631578947367, "percentage": 4.21, "elapsed_time": "0:00:11", "remaining_time": "0:04:22"}
5
+ {"current_steps": 50, "total_steps": 950, "loss": 0.7425, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.974346616959476e-05, "epoch": 0.2631578947368421, "percentage": 5.26, "elapsed_time": "0:00:14", "remaining_time": "0:04:14"}
6
+ {"current_steps": 60, "total_steps": 950, "loss": 0.7609, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.954424340791196e-05, "epoch": 0.3157894736842105, "percentage": 6.32, "elapsed_time": "0:00:16", "remaining_time": "0:04:07"}
7
+ {"current_steps": 70, "total_steps": 950, "loss": 0.6316, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.928848976574019e-05, "epoch": 0.3684210526315789, "percentage": 7.37, "elapsed_time": "0:00:19", "remaining_time": "0:04:01"}
8
+ {"current_steps": 80, "total_steps": 950, "loss": 0.6406, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.897649706262473e-05, "epoch": 0.42105263157894735, "percentage": 8.42, "elapsed_time": "0:00:21", "remaining_time": "0:03:56"}
9
+ {"current_steps": 90, "total_steps": 950, "loss": 0.5597, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.860862128789953e-05, "epoch": 0.47368421052631576, "percentage": 9.47, "elapsed_time": "0:00:24", "remaining_time": "0:03:52"}
10
+ {"current_steps": 100, "total_steps": 950, "loss": 0.477, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.818528219449705e-05, "epoch": 0.5263157894736842, "percentage": 10.53, "elapsed_time": "0:00:26", "remaining_time": "0:03:48"}
11
+ {"current_steps": 100, "total_steps": 950, "loss": null, "eval_loss": 0.46219402551651, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.5263157894736842, "percentage": 10.53, "elapsed_time": "0:00:26", "remaining_time": "0:03:48"}
12
+ {"current_steps": 110, "total_steps": 950, "loss": 0.4798, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.770696282000244e-05, "epoch": 0.5789473684210527, "percentage": 11.58, "elapsed_time": "0:00:30", "remaining_time": "0:03:53"}
13
+ {"current_steps": 120, "total_steps": 950, "loss": 0.5097, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.717420893549902e-05, "epoch": 0.631578947368421, "percentage": 12.63, "elapsed_time": "0:00:33", "remaining_time": "0:03:49"}
14
+ {"current_steps": 130, "total_steps": 950, "loss": 0.6507, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.658762842283343e-05, "epoch": 0.6842105263157895, "percentage": 13.68, "elapsed_time": "0:00:35", "remaining_time": "0:03:45"}
15
+ {"current_steps": 140, "total_steps": 950, "loss": 0.4961, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.594789058101153e-05, "epoch": 0.7368421052631579, "percentage": 14.74, "elapsed_time": "0:00:38", "remaining_time": "0:03:41"}
16
+ {"current_steps": 150, "total_steps": 950, "loss": 0.3643, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.525572536251607e-05, "epoch": 0.7894736842105263, "percentage": 15.79, "elapsed_time": "0:00:40", "remaining_time": "0:03:37"}
17
+ {"current_steps": 160, "total_steps": 950, "loss": 0.5859, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.451192254041758e-05, "epoch": 0.8421052631578947, "percentage": 16.84, "elapsed_time": "0:00:43", "remaining_time": "0:03:33"}
18
+ {"current_steps": 170, "total_steps": 950, "loss": 0.4765, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.371733080722911e-05, "epoch": 0.8947368421052632, "percentage": 17.89, "elapsed_time": "0:00:45", "remaining_time": "0:03:30"}
19
+ {"current_steps": 180, "total_steps": 950, "loss": 0.4496, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.287285680653254e-05, "epoch": 0.9473684210526315, "percentage": 18.95, "elapsed_time": "0:00:48", "remaining_time": "0:03:27"}
20
+ {"current_steps": 190, "total_steps": 950, "loss": 0.42, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.197946409848194e-05, "epoch": 1.0, "percentage": 20.0, "elapsed_time": "0:00:50", "remaining_time": "0:03:23"}
21
+ {"current_steps": 200, "total_steps": 950, "loss": 0.4647, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.103817206036382e-05, "epoch": 1.0526315789473684, "percentage": 21.05, "elapsed_time": "0:00:53", "remaining_time": "0:03:20"}
22
+ {"current_steps": 200, "total_steps": 950, "loss": null, "eval_loss": 0.2513052821159363, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.0526315789473684, "percentage": 21.05, "elapsed_time": "0:00:53", "remaining_time": "0:03:20"}
23
+ {"current_steps": 210, "total_steps": 950, "loss": 0.3552, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.005005472346924e-05, "epoch": 1.1052631578947367, "percentage": 22.11, "elapsed_time": "0:00:57", "remaining_time": "0:03:21"}
24
+ {"current_steps": 220, "total_steps": 950, "loss": 0.5985, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.90162395476046e-05, "epoch": 1.1578947368421053, "percentage": 23.16, "elapsed_time": "0:00:59", "remaining_time": "0:03:18"}
25
+ {"current_steps": 230, "total_steps": 950, "loss": 0.6665, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.793790613463955e-05, "epoch": 1.2105263157894737, "percentage": 24.21, "elapsed_time": "0:01:02", "remaining_time": "0:03:15"}
26
+ {"current_steps": 240, "total_steps": 950, "loss": 0.4249, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.681628488255986e-05, "epoch": 1.263157894736842, "percentage": 25.26, "elapsed_time": "0:01:04", "remaining_time": "0:03:11"}
27
+ {"current_steps": 250, "total_steps": 950, "loss": 0.4735, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.565265558156101e-05, "epoch": 1.3157894736842106, "percentage": 26.32, "elapsed_time": "0:01:07", "remaining_time": "0:03:08"}
28
+ {"current_steps": 260, "total_steps": 950, "loss": 0.265, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.444834595378434e-05, "epoch": 1.368421052631579, "percentage": 27.37, "elapsed_time": "0:01:09", "remaining_time": "0:03:05"}
29
+ {"current_steps": 270, "total_steps": 950, "loss": 0.5925, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.320473013836196e-05, "epoch": 1.4210526315789473, "percentage": 28.42, "elapsed_time": "0:01:12", "remaining_time": "0:03:02"}
30
+ {"current_steps": 280, "total_steps": 950, "loss": 0.4528, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.192322712349917e-05, "epoch": 1.4736842105263157, "percentage": 29.47, "elapsed_time": "0:01:14", "remaining_time": "0:02:59"}
31
+ {"current_steps": 290, "total_steps": 950, "loss": 0.3558, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.060529912738315e-05, "epoch": 1.526315789473684, "percentage": 30.53, "elapsed_time": "0:01:17", "remaining_time": "0:02:56"}
32
+ {"current_steps": 300, "total_steps": 950, "loss": 0.5231, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.925244992976538e-05, "epoch": 1.5789473684210527, "percentage": 31.58, "elapsed_time": "0:01:20", "remaining_time": "0:02:53"}
33
+ {"current_steps": 300, "total_steps": 950, "loss": null, "eval_loss": 0.34315773844718933, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.5789473684210527, "percentage": 31.58, "elapsed_time": "0:01:20", "remaining_time": "0:02:53"}
34
+ {"current_steps": 310, "total_steps": 950, "loss": 0.3751, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.786622315612183e-05, "epoch": 1.631578947368421, "percentage": 32.63, "elapsed_time": "0:01:23", "remaining_time": "0:02:52"}
35
+ {"current_steps": 320, "total_steps": 950, "loss": 0.3844, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.644820051634812e-05, "epoch": 1.6842105263157894, "percentage": 33.68, "elapsed_time": "0:01:26", "remaining_time": "0:02:49"}
36
+ {"current_steps": 330, "total_steps": 950, "loss": 0.5589, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.500000000000001e-05, "epoch": 1.736842105263158, "percentage": 34.74, "elapsed_time": "0:01:28", "remaining_time": "0:02:46"}
37
+ {"current_steps": 340, "total_steps": 950, "loss": 0.2688, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.35232740301378e-05, "epoch": 1.7894736842105263, "percentage": 35.79, "elapsed_time": "0:01:31", "remaining_time": "0:02:43"}
38
+ {"current_steps": 350, "total_steps": 950, "loss": 0.2528, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.201970757788172e-05, "epoch": 1.8421052631578947, "percentage": 36.84, "elapsed_time": "0:01:33", "remaining_time": "0:02:40"}
39
+ {"current_steps": 360, "total_steps": 950, "loss": 0.2398, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.049101623982937e-05, "epoch": 1.8947368421052633, "percentage": 37.89, "elapsed_time": "0:01:36", "remaining_time": "0:02:38"}
40
+ {"current_steps": 370, "total_steps": 950, "loss": 0.4979, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.89389442805288e-05, "epoch": 1.9473684210526314, "percentage": 38.95, "elapsed_time": "0:01:39", "remaining_time": "0:02:35"}
41
+ {"current_steps": 380, "total_steps": 950, "loss": 0.5478, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.736526264224101e-05, "epoch": 2.0, "percentage": 40.0, "elapsed_time": "0:01:41", "remaining_time": "0:02:32"}
42
+ {"current_steps": 390, "total_steps": 950, "loss": 0.4545, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.577176692426279e-05, "epoch": 2.0526315789473686, "percentage": 41.05, "elapsed_time": "0:01:44", "remaining_time": "0:02:29"}
43
+ {"current_steps": 400, "total_steps": 950, "loss": 0.4381, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.416027533411519e-05, "epoch": 2.1052631578947367, "percentage": 42.11, "elapsed_time": "0:01:46", "remaining_time": "0:02:26"}
44
+ {"current_steps": 400, "total_steps": 950, "loss": null, "eval_loss": 0.21073441207408905, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.1052631578947367, "percentage": 42.11, "elapsed_time": "0:01:46", "remaining_time": "0:02:26"}
45
+ {"current_steps": 410, "total_steps": 950, "loss": 0.2746, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.253262661293604e-05, "epoch": 2.1578947368421053, "percentage": 43.16, "elapsed_time": "0:01:50", "remaining_time": "0:02:25"}
46
+ {"current_steps": 420, "total_steps": 950, "loss": 0.2276, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.0890677937442574e-05, "epoch": 2.2105263157894735, "percentage": 44.21, "elapsed_time": "0:01:52", "remaining_time": "0:02:22"}
47
+ {"current_steps": 430, "total_steps": 950, "loss": 0.2474, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.923630280085948e-05, "epoch": 2.263157894736842, "percentage": 45.26, "elapsed_time": "0:01:55", "remaining_time": "0:02:19"}
48
+ {"current_steps": 440, "total_steps": 950, "loss": 0.2644, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.757138887522884e-05, "epoch": 2.3157894736842106, "percentage": 46.32, "elapsed_time": "0:01:57", "remaining_time": "0:02:16"}
49
+ {"current_steps": 450, "total_steps": 950, "loss": 0.3294, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.5897835857542317e-05, "epoch": 2.3684210526315788, "percentage": 47.37, "elapsed_time": "0:02:00", "remaining_time": "0:02:13"}
50
+ {"current_steps": 460, "total_steps": 950, "loss": 0.5553, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.4217553302152237e-05, "epoch": 2.4210526315789473, "percentage": 48.42, "elapsed_time": "0:02:02", "remaining_time": "0:02:10"}
51
+ {"current_steps": 470, "total_steps": 950, "loss": 0.1107, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.2532458441935636e-05, "epoch": 2.473684210526316, "percentage": 49.47, "elapsed_time": "0:02:05", "remaining_time": "0:02:08"}
52
+ {"current_steps": 480, "total_steps": 950, "loss": 0.2507, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.084447400069655e-05, "epoch": 2.526315789473684, "percentage": 50.53, "elapsed_time": "0:02:08", "remaining_time": "0:02:05"}
53
+ {"current_steps": 490, "total_steps": 950, "loss": 0.1839, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.915552599930345e-05, "epoch": 2.5789473684210527, "percentage": 51.58, "elapsed_time": "0:02:10", "remaining_time": "0:02:02"}
54
+ {"current_steps": 500, "total_steps": 950, "loss": 0.2159, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.746754155806437e-05, "epoch": 2.6315789473684212, "percentage": 52.63, "elapsed_time": "0:02:13", "remaining_time": "0:01:59"}
55
+ {"current_steps": 500, "total_steps": 950, "loss": null, "eval_loss": 0.36658158898353577, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.6315789473684212, "percentage": 52.63, "elapsed_time": "0:02:13", "remaining_time": "0:01:59"}
56
+ {"current_steps": 510, "total_steps": 950, "loss": 0.1742, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.578244669784777e-05, "epoch": 2.6842105263157894, "percentage": 53.68, "elapsed_time": "0:02:16", "remaining_time": "0:01:58"}
57
+ {"current_steps": 520, "total_steps": 950, "loss": 0.3915, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.410216414245771e-05, "epoch": 2.736842105263158, "percentage": 54.74, "elapsed_time": "0:02:19", "remaining_time": "0:01:55"}
58
+ {"current_steps": 530, "total_steps": 950, "loss": 0.2055, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.2428611124771184e-05, "epoch": 2.7894736842105265, "percentage": 55.79, "elapsed_time": "0:02:22", "remaining_time": "0:01:52"}
59
+ {"current_steps": 540, "total_steps": 950, "loss": 0.3395, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.076369719914055e-05, "epoch": 2.8421052631578947, "percentage": 56.84, "elapsed_time": "0:02:24", "remaining_time": "0:01:49"}
60
+ {"current_steps": 550, "total_steps": 950, "loss": 0.2351, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.9109322062557424e-05, "epoch": 2.8947368421052633, "percentage": 57.89, "elapsed_time": "0:02:27", "remaining_time": "0:01:47"}
61
+ {"current_steps": 560, "total_steps": 950, "loss": 0.1548, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.746737338706397e-05, "epoch": 2.9473684210526314, "percentage": 58.95, "elapsed_time": "0:02:29", "remaining_time": "0:01:44"}
62
+ {"current_steps": 570, "total_steps": 950, "loss": 0.1734, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.58397246658848e-05, "epoch": 3.0, "percentage": 60.0, "elapsed_time": "0:02:32", "remaining_time": "0:01:41"}
63
+ {"current_steps": 580, "total_steps": 950, "loss": 0.1015, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.422823307573722e-05, "epoch": 3.0526315789473686, "percentage": 61.05, "elapsed_time": "0:02:35", "remaining_time": "0:01:38"}
64
+ {"current_steps": 590, "total_steps": 950, "loss": 0.1705, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.263473735775899e-05, "epoch": 3.1052631578947367, "percentage": 62.11, "elapsed_time": "0:02:37", "remaining_time": "0:01:36"}
65
+ {"current_steps": 600, "total_steps": 950, "loss": 0.1934, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.10610557194712e-05, "epoch": 3.1578947368421053, "percentage": 63.16, "elapsed_time": "0:02:40", "remaining_time": "0:01:33"}
66
+ {"current_steps": 600, "total_steps": 950, "loss": null, "eval_loss": 0.26252931356430054, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.1578947368421053, "percentage": 63.16, "elapsed_time": "0:02:40", "remaining_time": "0:01:33"}
67
+ {"current_steps": 610, "total_steps": 950, "loss": 0.1024, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.950898376017064e-05, "epoch": 3.2105263157894735, "percentage": 64.21, "elapsed_time": "0:02:43", "remaining_time": "0:01:31"}
68
+ {"current_steps": 620, "total_steps": 950, "loss": 0.1786, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.798029242211828e-05, "epoch": 3.263157894736842, "percentage": 65.26, "elapsed_time": "0:02:46", "remaining_time": "0:01:28"}
69
+ {"current_steps": 630, "total_steps": 950, "loss": 0.1907, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.6476725969862227e-05, "epoch": 3.3157894736842106, "percentage": 66.32, "elapsed_time": "0:02:49", "remaining_time": "0:01:25"}
70
+ {"current_steps": 640, "total_steps": 950, "loss": 0.125, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.500000000000001e-05, "epoch": 3.3684210526315788, "percentage": 67.37, "elapsed_time": "0:02:51", "remaining_time": "0:01:23"}
71
+ {"current_steps": 650, "total_steps": 950, "loss": 0.1509, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.3551799483651894e-05, "epoch": 3.4210526315789473, "percentage": 68.42, "elapsed_time": "0:02:54", "remaining_time": "0:01:20"}
72
+ {"current_steps": 660, "total_steps": 950, "loss": 0.1502, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.2133776843878186e-05, "epoch": 3.473684210526316, "percentage": 69.47, "elapsed_time": "0:02:56", "remaining_time": "0:01:17"}
73
+ {"current_steps": 670, "total_steps": 950, "loss": 0.1458, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.074755007023461e-05, "epoch": 3.526315789473684, "percentage": 70.53, "elapsed_time": "0:02:59", "remaining_time": "0:01:14"}
74
+ {"current_steps": 680, "total_steps": 950, "loss": 0.2181, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9394700872616855e-05, "epoch": 3.5789473684210527, "percentage": 71.58, "elapsed_time": "0:03:01", "remaining_time": "0:01:12"}
75
+ {"current_steps": 690, "total_steps": 950, "loss": 0.0979, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.807677287650083e-05, "epoch": 3.6315789473684212, "percentage": 72.63, "elapsed_time": "0:03:04", "remaining_time": "0:01:09"}
76
+ {"current_steps": 700, "total_steps": 950, "loss": 0.2701, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.6795269861638042e-05, "epoch": 3.6842105263157894, "percentage": 73.68, "elapsed_time": "0:03:06", "remaining_time": "0:01:06"}
77
+ {"current_steps": 700, "total_steps": 950, "loss": null, "eval_loss": 0.2853814661502838, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.6842105263157894, "percentage": 73.68, "elapsed_time": "0:03:06", "remaining_time": "0:01:06"}
78
+ {"current_steps": 700, "total_steps": 950, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.6842105263157894, "percentage": 73.68, "elapsed_time": "0:03:06", "remaining_time": "0:01:06"}
79
+ {"current_steps": 10, "total_steps": 10, "loss": null, "eval_loss": 0.21073441207408905, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.6842105263157894, "percentage": 100.0, "elapsed_time": "0:03:09", "remaining_time": "0:00:00"}
llama2_13b_peft/object_counting/trainer_state.json ADDED
@@ -0,0 +1,576 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.21073441207408905,
3
+ "best_model_checkpoint": "ckpt/llama2_13b_fuze27_no_sys/object_counting_no_sys/checkpoint-400",
4
+ "epoch": 3.6842105263157894,
5
+ "eval_steps": 100,
6
+ "global_step": 700,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.05263157894736842,
13
+ "grad_norm": 6.26442289352417,
14
+ "learning_rate": 5e-05,
15
+ "loss": 10.645,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.10526315789473684,
20
+ "grad_norm": 4.90908145904541,
21
+ "learning_rate": 0.0001,
22
+ "loss": 4.9002,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.15789473684210525,
27
+ "grad_norm": 7.5457682609558105,
28
+ "learning_rate": 9.99714745464859e-05,
29
+ "loss": 0.9012,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.21052631578947367,
34
+ "grad_norm": 9.139291763305664,
35
+ "learning_rate": 9.988593073400354e-05,
36
+ "loss": 0.7022,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.2631578947368421,
41
+ "grad_norm": 5.00139856338501,
42
+ "learning_rate": 9.974346616959476e-05,
43
+ "loss": 0.7425,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.3157894736842105,
48
+ "grad_norm": 6.002871990203857,
49
+ "learning_rate": 9.954424340791196e-05,
50
+ "loss": 0.7609,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.3684210526315789,
55
+ "grad_norm": 6.101377010345459,
56
+ "learning_rate": 9.928848976574019e-05,
57
+ "loss": 0.6316,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.42105263157894735,
62
+ "grad_norm": 5.7651824951171875,
63
+ "learning_rate": 9.897649706262473e-05,
64
+ "loss": 0.6406,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.47368421052631576,
69
+ "grad_norm": 2.46134614944458,
70
+ "learning_rate": 9.860862128789953e-05,
71
+ "loss": 0.5597,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.5263157894736842,
76
+ "grad_norm": 4.3927178382873535,
77
+ "learning_rate": 9.818528219449705e-05,
78
+ "loss": 0.477,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.5263157894736842,
83
+ "eval_loss": 0.46219402551651,
84
+ "eval_runtime": 0.6936,
85
+ "eval_samples_per_second": 57.669,
86
+ "eval_steps_per_second": 14.417,
87
+ "step": 100
88
+ },
89
+ {
90
+ "epoch": 0.5789473684210527,
91
+ "grad_norm": 6.058330059051514,
92
+ "learning_rate": 9.770696282000244e-05,
93
+ "loss": 0.4798,
94
+ "step": 110
95
+ },
96
+ {
97
+ "epoch": 0.631578947368421,
98
+ "grad_norm": 4.410764217376709,
99
+ "learning_rate": 9.717420893549902e-05,
100
+ "loss": 0.5097,
101
+ "step": 120
102
+ },
103
+ {
104
+ "epoch": 0.6842105263157895,
105
+ "grad_norm": 1.7051620483398438,
106
+ "learning_rate": 9.658762842283343e-05,
107
+ "loss": 0.6507,
108
+ "step": 130
109
+ },
110
+ {
111
+ "epoch": 0.7368421052631579,
112
+ "grad_norm": 3.638519763946533,
113
+ "learning_rate": 9.594789058101153e-05,
114
+ "loss": 0.4961,
115
+ "step": 140
116
+ },
117
+ {
118
+ "epoch": 0.7894736842105263,
119
+ "grad_norm": 1.8080860376358032,
120
+ "learning_rate": 9.525572536251607e-05,
121
+ "loss": 0.3643,
122
+ "step": 150
123
+ },
124
+ {
125
+ "epoch": 0.8421052631578947,
126
+ "grad_norm": 4.843872547149658,
127
+ "learning_rate": 9.451192254041758e-05,
128
+ "loss": 0.5859,
129
+ "step": 160
130
+ },
131
+ {
132
+ "epoch": 0.8947368421052632,
133
+ "grad_norm": 7.292293548583984,
134
+ "learning_rate": 9.371733080722911e-05,
135
+ "loss": 0.4765,
136
+ "step": 170
137
+ },
138
+ {
139
+ "epoch": 0.9473684210526315,
140
+ "grad_norm": 2.1483707427978516,
141
+ "learning_rate": 9.287285680653254e-05,
142
+ "loss": 0.4496,
143
+ "step": 180
144
+ },
145
+ {
146
+ "epoch": 1.0,
147
+ "grad_norm": 1.2343636751174927,
148
+ "learning_rate": 9.197946409848194e-05,
149
+ "loss": 0.42,
150
+ "step": 190
151
+ },
152
+ {
153
+ "epoch": 1.0526315789473684,
154
+ "grad_norm": 3.902963638305664,
155
+ "learning_rate": 9.103817206036382e-05,
156
+ "loss": 0.4647,
157
+ "step": 200
158
+ },
159
+ {
160
+ "epoch": 1.0526315789473684,
161
+ "eval_loss": 0.2513052821159363,
162
+ "eval_runtime": 0.6934,
163
+ "eval_samples_per_second": 57.69,
164
+ "eval_steps_per_second": 14.423,
165
+ "step": 200
166
+ },
167
+ {
168
+ "epoch": 1.1052631578947367,
169
+ "grad_norm": 3.0811166763305664,
170
+ "learning_rate": 9.005005472346924e-05,
171
+ "loss": 0.3552,
172
+ "step": 210
173
+ },
174
+ {
175
+ "epoch": 1.1578947368421053,
176
+ "grad_norm": 12.97741413116455,
177
+ "learning_rate": 8.90162395476046e-05,
178
+ "loss": 0.5985,
179
+ "step": 220
180
+ },
181
+ {
182
+ "epoch": 1.2105263157894737,
183
+ "grad_norm": 4.255702495574951,
184
+ "learning_rate": 8.793790613463955e-05,
185
+ "loss": 0.6665,
186
+ "step": 230
187
+ },
188
+ {
189
+ "epoch": 1.263157894736842,
190
+ "grad_norm": 3.5546047687530518,
191
+ "learning_rate": 8.681628488255986e-05,
192
+ "loss": 0.4249,
193
+ "step": 240
194
+ },
195
+ {
196
+ "epoch": 1.3157894736842106,
197
+ "grad_norm": 9.395575523376465,
198
+ "learning_rate": 8.565265558156101e-05,
199
+ "loss": 0.4735,
200
+ "step": 250
201
+ },
202
+ {
203
+ "epoch": 1.368421052631579,
204
+ "grad_norm": 3.7864913940429688,
205
+ "learning_rate": 8.444834595378434e-05,
206
+ "loss": 0.265,
207
+ "step": 260
208
+ },
209
+ {
210
+ "epoch": 1.4210526315789473,
211
+ "grad_norm": 6.043397426605225,
212
+ "learning_rate": 8.320473013836196e-05,
213
+ "loss": 0.5925,
214
+ "step": 270
215
+ },
216
+ {
217
+ "epoch": 1.4736842105263157,
218
+ "grad_norm": 5.700586795806885,
219
+ "learning_rate": 8.192322712349917e-05,
220
+ "loss": 0.4528,
221
+ "step": 280
222
+ },
223
+ {
224
+ "epoch": 1.526315789473684,
225
+ "grad_norm": 2.9391582012176514,
226
+ "learning_rate": 8.060529912738315e-05,
227
+ "loss": 0.3558,
228
+ "step": 290
229
+ },
230
+ {
231
+ "epoch": 1.5789473684210527,
232
+ "grad_norm": 1.8771816492080688,
233
+ "learning_rate": 7.925244992976538e-05,
234
+ "loss": 0.5231,
235
+ "step": 300
236
+ },
237
+ {
238
+ "epoch": 1.5789473684210527,
239
+ "eval_loss": 0.34315773844718933,
240
+ "eval_runtime": 0.6944,
241
+ "eval_samples_per_second": 57.607,
242
+ "eval_steps_per_second": 14.402,
243
+ "step": 300
244
+ },
245
+ {
246
+ "epoch": 1.631578947368421,
247
+ "grad_norm": 4.8421478271484375,
248
+ "learning_rate": 7.786622315612183e-05,
249
+ "loss": 0.3751,
250
+ "step": 310
251
+ },
252
+ {
253
+ "epoch": 1.6842105263157894,
254
+ "grad_norm": 11.370196342468262,
255
+ "learning_rate": 7.644820051634812e-05,
256
+ "loss": 0.3844,
257
+ "step": 320
258
+ },
259
+ {
260
+ "epoch": 1.736842105263158,
261
+ "grad_norm": 3.433976173400879,
262
+ "learning_rate": 7.500000000000001e-05,
263
+ "loss": 0.5589,
264
+ "step": 330
265
+ },
266
+ {
267
+ "epoch": 1.7894736842105263,
268
+ "grad_norm": 6.323594570159912,
269
+ "learning_rate": 7.35232740301378e-05,
270
+ "loss": 0.2688,
271
+ "step": 340
272
+ },
273
+ {
274
+ "epoch": 1.8421052631578947,
275
+ "grad_norm": 0.5850329399108887,
276
+ "learning_rate": 7.201970757788172e-05,
277
+ "loss": 0.2528,
278
+ "step": 350
279
+ },
280
+ {
281
+ "epoch": 1.8947368421052633,
282
+ "grad_norm": 1.9679738283157349,
283
+ "learning_rate": 7.049101623982937e-05,
284
+ "loss": 0.2398,
285
+ "step": 360
286
+ },
287
+ {
288
+ "epoch": 1.9473684210526314,
289
+ "grad_norm": 12.53429889678955,
290
+ "learning_rate": 6.89389442805288e-05,
291
+ "loss": 0.4979,
292
+ "step": 370
293
+ },
294
+ {
295
+ "epoch": 2.0,
296
+ "grad_norm": 7.08478307723999,
297
+ "learning_rate": 6.736526264224101e-05,
298
+ "loss": 0.5478,
299
+ "step": 380
300
+ },
301
+ {
302
+ "epoch": 2.0526315789473686,
303
+ "grad_norm": 9.115711212158203,
304
+ "learning_rate": 6.577176692426279e-05,
305
+ "loss": 0.4545,
306
+ "step": 390
307
+ },
308
+ {
309
+ "epoch": 2.1052631578947367,
310
+ "grad_norm": 6.18236780166626,
311
+ "learning_rate": 6.416027533411519e-05,
312
+ "loss": 0.4381,
313
+ "step": 400
314
+ },
315
+ {
316
+ "epoch": 2.1052631578947367,
317
+ "eval_loss": 0.21073441207408905,
318
+ "eval_runtime": 0.6912,
319
+ "eval_samples_per_second": 57.873,
320
+ "eval_steps_per_second": 14.468,
321
+ "step": 400
322
+ },
323
+ {
324
+ "epoch": 2.1578947368421053,
325
+ "grad_norm": 1.294372797012329,
326
+ "learning_rate": 6.253262661293604e-05,
327
+ "loss": 0.2746,
328
+ "step": 410
329
+ },
330
+ {
331
+ "epoch": 2.2105263157894735,
332
+ "grad_norm": 9.00351619720459,
333
+ "learning_rate": 6.0890677937442574e-05,
334
+ "loss": 0.2276,
335
+ "step": 420
336
+ },
337
+ {
338
+ "epoch": 2.263157894736842,
339
+ "grad_norm": 10.986132621765137,
340
+ "learning_rate": 5.923630280085948e-05,
341
+ "loss": 0.2474,
342
+ "step": 430
343
+ },
344
+ {
345
+ "epoch": 2.3157894736842106,
346
+ "grad_norm": 12.219366073608398,
347
+ "learning_rate": 5.757138887522884e-05,
348
+ "loss": 0.2644,
349
+ "step": 440
350
+ },
351
+ {
352
+ "epoch": 2.3684210526315788,
353
+ "grad_norm": 2.5177173614501953,
354
+ "learning_rate": 5.5897835857542317e-05,
355
+ "loss": 0.3294,
356
+ "step": 450
357
+ },
358
+ {
359
+ "epoch": 2.4210526315789473,
360
+ "grad_norm": 6.946788787841797,
361
+ "learning_rate": 5.4217553302152237e-05,
362
+ "loss": 0.5553,
363
+ "step": 460
364
+ },
365
+ {
366
+ "epoch": 2.473684210526316,
367
+ "grad_norm": 3.0805630683898926,
368
+ "learning_rate": 5.2532458441935636e-05,
369
+ "loss": 0.1107,
370
+ "step": 470
371
+ },
372
+ {
373
+ "epoch": 2.526315789473684,
374
+ "grad_norm": 3.5907771587371826,
375
+ "learning_rate": 5.084447400069655e-05,
376
+ "loss": 0.2507,
377
+ "step": 480
378
+ },
379
+ {
380
+ "epoch": 2.5789473684210527,
381
+ "grad_norm": 1.5780644416809082,
382
+ "learning_rate": 4.915552599930345e-05,
383
+ "loss": 0.1839,
384
+ "step": 490
385
+ },
386
+ {
387
+ "epoch": 2.6315789473684212,
388
+ "grad_norm": 2.739733934402466,
389
+ "learning_rate": 4.746754155806437e-05,
390
+ "loss": 0.2159,
391
+ "step": 500
392
+ },
393
+ {
394
+ "epoch": 2.6315789473684212,
395
+ "eval_loss": 0.36658158898353577,
396
+ "eval_runtime": 0.6913,
397
+ "eval_samples_per_second": 57.864,
398
+ "eval_steps_per_second": 14.466,
399
+ "step": 500
400
+ },
401
+ {
402
+ "epoch": 2.6842105263157894,
403
+ "grad_norm": 1.3136796951293945,
404
+ "learning_rate": 4.578244669784777e-05,
405
+ "loss": 0.1742,
406
+ "step": 510
407
+ },
408
+ {
409
+ "epoch": 2.736842105263158,
410
+ "grad_norm": 0.0468708761036396,
411
+ "learning_rate": 4.410216414245771e-05,
412
+ "loss": 0.3915,
413
+ "step": 520
414
+ },
415
+ {
416
+ "epoch": 2.7894736842105265,
417
+ "grad_norm": 12.833720207214355,
418
+ "learning_rate": 4.2428611124771184e-05,
419
+ "loss": 0.2055,
420
+ "step": 530
421
+ },
422
+ {
423
+ "epoch": 2.8421052631578947,
424
+ "grad_norm": 0.10741530358791351,
425
+ "learning_rate": 4.076369719914055e-05,
426
+ "loss": 0.3395,
427
+ "step": 540
428
+ },
429
+ {
430
+ "epoch": 2.8947368421052633,
431
+ "grad_norm": 6.424005508422852,
432
+ "learning_rate": 3.9109322062557424e-05,
433
+ "loss": 0.2351,
434
+ "step": 550
435
+ },
436
+ {
437
+ "epoch": 2.9473684210526314,
438
+ "grad_norm": 7.805177211761475,
439
+ "learning_rate": 3.746737338706397e-05,
440
+ "loss": 0.1548,
441
+ "step": 560
442
+ },
443
+ {
444
+ "epoch": 3.0,
445
+ "grad_norm": 1.9612077474594116,
446
+ "learning_rate": 3.58397246658848e-05,
447
+ "loss": 0.1734,
448
+ "step": 570
449
+ },
450
+ {
451
+ "epoch": 3.0526315789473686,
452
+ "grad_norm": 6.020046234130859,
453
+ "learning_rate": 3.422823307573722e-05,
454
+ "loss": 0.1015,
455
+ "step": 580
456
+ },
457
+ {
458
+ "epoch": 3.1052631578947367,
459
+ "grad_norm": 5.7347307205200195,
460
+ "learning_rate": 3.263473735775899e-05,
461
+ "loss": 0.1705,
462
+ "step": 590
463
+ },
464
+ {
465
+ "epoch": 3.1578947368421053,
466
+ "grad_norm": 15.45874309539795,
467
+ "learning_rate": 3.10610557194712e-05,
468
+ "loss": 0.1934,
469
+ "step": 600
470
+ },
471
+ {
472
+ "epoch": 3.1578947368421053,
473
+ "eval_loss": 0.26252931356430054,
474
+ "eval_runtime": 0.6908,
475
+ "eval_samples_per_second": 57.906,
476
+ "eval_steps_per_second": 14.476,
477
+ "step": 600
478
+ },
479
+ {
480
+ "epoch": 3.2105263157894735,
481
+ "grad_norm": 1.443434715270996,
482
+ "learning_rate": 2.950898376017064e-05,
483
+ "loss": 0.1024,
484
+ "step": 610
485
+ },
486
+ {
487
+ "epoch": 3.263157894736842,
488
+ "grad_norm": 8.99201488494873,
489
+ "learning_rate": 2.798029242211828e-05,
490
+ "loss": 0.1786,
491
+ "step": 620
492
+ },
493
+ {
494
+ "epoch": 3.3157894736842106,
495
+ "grad_norm": 3.1811983585357666,
496
+ "learning_rate": 2.6476725969862227e-05,
497
+ "loss": 0.1907,
498
+ "step": 630
499
+ },
500
+ {
501
+ "epoch": 3.3684210526315788,
502
+ "grad_norm": 16.603193283081055,
503
+ "learning_rate": 2.500000000000001e-05,
504
+ "loss": 0.125,
505
+ "step": 640
506
+ },
507
+ {
508
+ "epoch": 3.4210526315789473,
509
+ "grad_norm": 6.346674919128418,
510
+ "learning_rate": 2.3551799483651894e-05,
511
+ "loss": 0.1509,
512
+ "step": 650
513
+ },
514
+ {
515
+ "epoch": 3.473684210526316,
516
+ "grad_norm": 3.2951841354370117,
517
+ "learning_rate": 2.2133776843878186e-05,
518
+ "loss": 0.1502,
519
+ "step": 660
520
+ },
521
+ {
522
+ "epoch": 3.526315789473684,
523
+ "grad_norm": 4.906289100646973,
524
+ "learning_rate": 2.074755007023461e-05,
525
+ "loss": 0.1458,
526
+ "step": 670
527
+ },
528
+ {
529
+ "epoch": 3.5789473684210527,
530
+ "grad_norm": 0.9275074005126953,
531
+ "learning_rate": 1.9394700872616855e-05,
532
+ "loss": 0.2181,
533
+ "step": 680
534
+ },
535
+ {
536
+ "epoch": 3.6315789473684212,
537
+ "grad_norm": 0.8629674315452576,
538
+ "learning_rate": 1.807677287650083e-05,
539
+ "loss": 0.0979,
540
+ "step": 690
541
+ },
542
+ {
543
+ "epoch": 3.6842105263157894,
544
+ "grad_norm": 6.139326095581055,
545
+ "learning_rate": 1.6795269861638042e-05,
546
+ "loss": 0.2701,
547
+ "step": 700
548
+ },
549
+ {
550
+ "epoch": 3.6842105263157894,
551
+ "eval_loss": 0.2853814661502838,
552
+ "eval_runtime": 0.691,
553
+ "eval_samples_per_second": 57.884,
554
+ "eval_steps_per_second": 14.471,
555
+ "step": 700
556
+ },
557
+ {
558
+ "epoch": 3.6842105263157894,
559
+ "step": 700,
560
+ "total_flos": 1.461153753759744e+16,
561
+ "train_loss": 0.5859067852156503,
562
+ "train_runtime": 192.1858,
563
+ "train_samples_per_second": 19.773,
564
+ "train_steps_per_second": 4.943
565
+ }
566
+ ],
567
+ "logging_steps": 10,
568
+ "max_steps": 950,
569
+ "num_input_tokens_seen": 0,
570
+ "num_train_epochs": 5,
571
+ "save_steps": 100,
572
+ "total_flos": 1.461153753759744e+16,
573
+ "train_batch_size": 4,
574
+ "trial_name": null,
575
+ "trial_params": null
576
+ }
llama2_13b_peft/object_counting/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84748a17ab24fc05e783a2f81ad62b8bc6f4c5abe17d42454e72ec6964e7435f
3
+ size 5176
llama2_13b_peft/object_counting/training_eval_loss.png ADDED
llama2_13b_peft/object_counting/training_loss.png ADDED