timewanderer commited on
Commit
54ffa57
1 Parent(s): 61db053

Training in progress, step 500

Browse files
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:819971911a78df84505868dabc1ffcc3e1e8a44041388f9e4bd02b17a871f385
3
  size 268290900
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3106101d76ba51ae6ac38bd5fbbb5381ba03142f77f63a4407420d2d5bd11364
3
  size 268290900
run-1/checkpoint-500/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:294322a3921efabcf477dc40b7b403c369a45a4957c3e80807fc97e64c4a09cf
3
  size 268290900
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3106101d76ba51ae6ac38bd5fbbb5381ba03142f77f63a4407420d2d5bd11364
3
  size 268290900
run-1/checkpoint-500/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5c90eb90b8269b7374850d7f98358b4fcfca644626c4d7e9cece862afe8f3803
3
  size 536643898
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a83150038ed4ba83e5bf33c67e3270163d41f46616194298f99d3d06c85a88a
3
  size 536643898
run-1/checkpoint-500/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9e1264523e958cf7990dc5f42d876cc12129475c4603804cf66868aaf25c2c24
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04366f62f8f88f5a8265df59adb051b320463277845db80e7fa43f13110c18c9
3
  size 1064
run-1/checkpoint-500/trainer_state.json CHANGED
@@ -10,25 +10,25 @@
10
  "log_history": [
11
  {
12
  "epoch": 1.0,
13
- "eval_accuracy": 0.6690322580645162,
14
- "eval_loss": 0.42948082089424133,
15
- "eval_runtime": 5.393,
16
- "eval_samples_per_second": 574.82,
17
- "eval_steps_per_second": 12.053,
18
  "step": 318
19
  },
20
  {
21
  "epoch": 1.5723270440251573,
22
- "grad_norm": 0.8646999001502991,
23
- "learning_rate": 1.371069182389937e-05,
24
- "loss": 0.6591,
25
  "step": 500
26
  }
27
  ],
28
  "logging_steps": 500,
29
- "max_steps": 1590,
30
  "num_input_tokens_seen": 0,
31
- "num_train_epochs": 5,
32
  "save_steps": 500,
33
  "stateful_callbacks": {
34
  "TrainerControl": {
@@ -46,8 +46,8 @@
46
  "train_batch_size": 48,
47
  "trial_name": null,
48
  "trial_params": {
49
- "alpha": 0.4868927644085208,
50
- "num_train_epochs": 5,
51
- "temperature": 2
52
  }
53
  }
 
10
  "log_history": [
11
  {
12
  "epoch": 1.0,
13
+ "eval_accuracy": 0.6529032258064517,
14
+ "eval_loss": 0.28555235266685486,
15
+ "eval_runtime": 5.4121,
16
+ "eval_samples_per_second": 572.786,
17
+ "eval_steps_per_second": 12.01,
18
  "step": 318
19
  },
20
  {
21
  "epoch": 1.5723270440251573,
22
+ "grad_norm": 0.6621501445770264,
23
+ "learning_rate": 1.685534591194969e-05,
24
+ "loss": 0.4579,
25
  "step": 500
26
  }
27
  ],
28
  "logging_steps": 500,
29
+ "max_steps": 3180,
30
  "num_input_tokens_seen": 0,
31
+ "num_train_epochs": 10,
32
  "save_steps": 500,
33
  "stateful_callbacks": {
34
  "TrainerControl": {
 
46
  "train_batch_size": 48,
47
  "trial_name": null,
48
  "trial_params": {
49
+ "alpha": 0.31688267813599835,
50
+ "num_train_epochs": 10,
51
+ "temperature": 3
52
  }
53
  }
run-1/checkpoint-500/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2f061e43c65f095a686a86278f34a31824db2598d218eedeec610b734a0e65f9
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a074013ebca87018fe538250c2f45bfdeb260149a400a0644508c302966c2ef
3
  size 5240
runs/Oct12_06-40-39_b76c1be2ae55/events.out.tfevents.1728716534.b76c1be2ae55.1423.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b71460b0071f46213a0848524a678d109de6f766c17d002a989004287e7f0698
3
+ size 13299
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0589ed65fdaae56362b9a8760696297fdd9596b2ddfb3d7051a76a6dfd129716
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a074013ebca87018fe538250c2f45bfdeb260149a400a0644508c302966c2ef
3
  size 5240