marialvsantiago commited on
Commit
5fcee43
·
verified ·
1 Parent(s): e61d940

Training in progress, step 20, checkpoint

Browse files
last-checkpoint/adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5cc3e7552b67fed184376e047b408e2b90cd43208998fdf397ed4ed018363ba7
3
  size 9450352
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bd0ef2d6f6fcc18bdffcfdd397ff7c2e1886149604070ddb67bbeaf2f51781f
3
  size 9450352
last-checkpoint/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a19a9760899d24c7773cca0837f28285832af790f8ea1bf046a9ebb602539c8
3
  size 18955258
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c9cfd1c8be71a7dc7b5a9bf4f9e1e89f3ea05538ab59788bf155fb541f91bc1
3
  size 18955258
last-checkpoint/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6db6ad40cf0e9cf794015d421853a3e1fcab6923e8a60784b5fb9bd91b69507f
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7360faabbbf2d89f1d5cd92a24544edfb7ce0c21ab7da12aa9849741b6aaf9c
3
  size 14244
last-checkpoint/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c188a6a4749e6ca627bb6d536eb7443f499d5b1b88d98a78f9c713443e010d9c
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:015707cb16790250630febca682498cb5d3456d5a13443b953687f19dc7d59ed
3
  size 1064
last-checkpoint/trainer_state.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
- "best_metric": 3.5034565925598145,
3
- "best_model_checkpoint": "miner_id_24/checkpoint-10",
4
- "epoch": 0.0006797403391904293,
5
  "eval_steps": 5,
6
- "global_step": 10,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
@@ -52,6 +52,43 @@
52
  "eval_samples_per_second": 75.116,
53
  "eval_steps_per_second": 18.783,
54
  "step": 10
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
55
  }
56
  ],
57
  "logging_steps": 3,
@@ -80,7 +117,7 @@
80
  "attributes": {}
81
  }
82
  },
83
- "total_flos": 127008755417088.0,
84
  "train_batch_size": 4,
85
  "trial_name": null,
86
  "trial_params": null
 
1
  {
2
+ "best_metric": 3.379162311553955,
3
+ "best_model_checkpoint": "miner_id_24/checkpoint-20",
4
+ "epoch": 0.0013594806783808586,
5
  "eval_steps": 5,
6
+ "global_step": 20,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
 
52
  "eval_samples_per_second": 75.116,
53
  "eval_steps_per_second": 18.783,
54
  "step": 10
55
+ },
56
+ {
57
+ "epoch": 0.0008156884070285151,
58
+ "grad_norm": 430.5657653808594,
59
+ "learning_rate": 0.000163742398974869,
60
+ "loss": 13.4003,
61
+ "step": 12
62
+ },
63
+ {
64
+ "epoch": 0.0010196105087856438,
65
+ "grad_norm": 397.2607116699219,
66
+ "learning_rate": 0.00013090169943749476,
67
+ "loss": 13.5369,
68
+ "step": 15
69
+ },
70
+ {
71
+ "epoch": 0.0010196105087856438,
72
+ "eval_loss": 3.441890239715576,
73
+ "eval_runtime": 164.7776,
74
+ "eval_samples_per_second": 75.186,
75
+ "eval_steps_per_second": 18.801,
76
+ "step": 15
77
+ },
78
+ {
79
+ "epoch": 0.0012235326105427726,
80
+ "grad_norm": 462.4978942871094,
81
+ "learning_rate": 9.372094804706867e-05,
82
+ "loss": 13.3164,
83
+ "step": 18
84
+ },
85
+ {
86
+ "epoch": 0.0013594806783808586,
87
+ "eval_loss": 3.379162311553955,
88
+ "eval_runtime": 164.6217,
89
+ "eval_samples_per_second": 75.257,
90
+ "eval_steps_per_second": 18.819,
91
+ "step": 20
92
  }
93
  ],
94
  "logging_steps": 3,
 
117
  "attributes": {}
118
  }
119
  },
120
+ "total_flos": 250919736311808.0,
121
  "train_batch_size": 4,
122
  "trial_name": null,
123
  "trial_params": null