dimasik87 commited on
Commit
3b43766
·
verified ·
1 Parent(s): 0f17fd7

Training in progress, step 30, checkpoint

Browse files
last-checkpoint/adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:98197836c9eccbf127f4b122d048b6aa09a2ddf7c7950af4c893b8fee8653fe4
3
  size 167832240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:231a5cf4d496a6afa8cc1409eca8d9a5df86e48d16a4f7172925a55d014ff785
3
  size 167832240
last-checkpoint/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c45d97042aa5b14e89e9c54779a26ad4bdef00259c260eeb6e1db7133457a6ac
3
  size 335922386
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a1bb7c2559a32edecaa1575dc87e1b10b1a50272049171beccb84eb70201bbb
3
  size 335922386
last-checkpoint/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6aa91133351808aacad3e45f05aebdf1a2981ac4aa08431c95904a973b1b12d5
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62ed3fd3312f96c17ea93ab3f7344a9def6141fa3d024e167875cba869f24ea5
3
  size 14244
last-checkpoint/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:015707cb16790250630febca682498cb5d3456d5a13443b953687f19dc7d59ed
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2def2cd24154d8cecbaa07c36ae27e5ebb9b7273a78abfea27aa67c480e4ae2b
3
  size 1064
last-checkpoint/trainer_state.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
- "best_metric": 1.8700758218765259,
3
- "best_model_checkpoint": "miner_id_24/checkpoint-20",
4
- "epoch": 0.0010486989578554107,
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,
@@ -89,6 +89,50 @@
89
  "eval_samples_per_second": 17.323,
90
  "eval_steps_per_second": 8.662,
91
  "step": 20
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
92
  }
93
  ],
94
  "logging_steps": 3,
@@ -112,12 +156,12 @@
112
  "should_evaluate": false,
113
  "should_log": false,
114
  "should_save": true,
115
- "should_training_stop": false
116
  },
117
  "attributes": {}
118
  }
119
  },
120
- "total_flos": 7418872835604480.0,
121
  "train_batch_size": 2,
122
  "trial_name": null,
123
  "trial_params": null
 
1
  {
2
+ "best_metric": 1.8558446168899536,
3
+ "best_model_checkpoint": "miner_id_24/checkpoint-30",
4
+ "epoch": 0.001573048436783116,
5
  "eval_steps": 5,
6
+ "global_step": 30,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
 
89
  "eval_samples_per_second": 17.323,
90
  "eval_steps_per_second": 8.662,
91
  "step": 20
92
+ },
93
+ {
94
+ "epoch": 0.0011011339057481811,
95
+ "grad_norm": 0.9151540994644165,
96
+ "learning_rate": 5.7422070843492734e-05,
97
+ "loss": 1.8715,
98
+ "step": 21
99
+ },
100
+ {
101
+ "epoch": 0.0012584387494264928,
102
+ "grad_norm": 0.7925032377243042,
103
+ "learning_rate": 2.7103137257858868e-05,
104
+ "loss": 1.7962,
105
+ "step": 24
106
+ },
107
+ {
108
+ "epoch": 0.0013108736973192633,
109
+ "eval_loss": 1.860071063041687,
110
+ "eval_runtime": 463.3944,
111
+ "eval_samples_per_second": 17.329,
112
+ "eval_steps_per_second": 8.664,
113
+ "step": 25
114
+ },
115
+ {
116
+ "epoch": 0.0014157435931048043,
117
+ "grad_norm": 0.8621093034744263,
118
+ "learning_rate": 7.022351411174866e-06,
119
+ "loss": 1.7846,
120
+ "step": 27
121
+ },
122
+ {
123
+ "epoch": 0.001573048436783116,
124
+ "grad_norm": 1.050020694732666,
125
+ "learning_rate": 0.0,
126
+ "loss": 1.8138,
127
+ "step": 30
128
+ },
129
+ {
130
+ "epoch": 0.001573048436783116,
131
+ "eval_loss": 1.8558446168899536,
132
+ "eval_runtime": 463.1752,
133
+ "eval_samples_per_second": 17.337,
134
+ "eval_steps_per_second": 8.668,
135
+ "step": 30
136
  }
137
  ],
138
  "logging_steps": 3,
 
156
  "should_evaluate": false,
157
  "should_log": false,
158
  "should_save": true,
159
+ "should_training_stop": true
160
  },
161
  "attributes": {}
162
  }
163
  },
164
+ "total_flos": 1.1221045163851776e+16,
165
  "train_batch_size": 2,
166
  "trial_name": null,
167
  "trial_params": null