timewanderer commited on
Commit
5a0552b
1 Parent(s): 1ad70bc

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 sha256:9a461fedb48bd3ba08cf0db08a34f253e9782b2d81e15c6fa47b4f754bb7a3c2
3
  size 268290900
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:294322a3921efabcf477dc40b7b403c369a45a4957c3e80807fc97e64c4a09cf
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 sha256:52e6108bc9c45049ab11c8fdd2fbe2bcab1ff27c482e767f4d1e92257fa1d9d8
3
  size 268290900
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:294322a3921efabcf477dc40b7b403c369a45a4957c3e80807fc97e64c4a09cf
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 sha256:313f092c84462d4fa7ee8138e4a4dae83dd47ede60153b0efa5fe63374f265d4
3
  size 536643898
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5c90eb90b8269b7374850d7f98358b4fcfca644626c4d7e9cece862afe8f3803
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 sha256:90dc4637e972cc69c745eebddd8a7560dca27d2318df3e23f8e145abbf236536
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e1264523e958cf7990dc5f42d876cc12129475c4603804cf66868aaf25c2c24
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": 0.5829032258064516,
14
- "eval_loss": 0.1985393762588501,
15
- "eval_runtime": 5.6203,
16
- "eval_samples_per_second": 551.572,
17
- "eval_steps_per_second": 11.565,
18
  "step": 318
19
  },
20
  {
21
  "epoch": 1.5723270440251573,
22
- "grad_norm": 0.5288612246513367,
23
- "learning_rate": 1.606918238993711e-05,
24
- "loss": 0.3184,
25
  "step": 500
26
  }
27
  ],
28
  "logging_steps": 500,
29
- "max_steps": 2544,
30
  "num_input_tokens_seen": 0,
31
- "num_train_epochs": 8,
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": 0.819089504077056,
50
- "num_train_epochs": 8,
51
- "temperature": 12
52
  }
53
  }
 
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
  "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
  }
run-1/checkpoint-500/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3ad3761ee6c049b5f54cddb21d4403152bfaafdbfc2a13b71993419e99660838
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f061e43c65f095a686a86278f34a31824db2598d218eedeec610b734a0e65f9
3
  size 5240
runs/Oct12_05-09-37_f46fcd247c16/events.out.tfevents.1728710945.f46fcd247c16.522.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e201aea418d11654f13843aa71bb6b798584937056d55d8996ff28ae2f3190ad
3
+ size 13298
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:30bdc23a7d8fb386178e976baa34d55beb56c6ffba63fe636be670468d00efc1
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f061e43c65f095a686a86278f34a31824db2598d218eedeec610b734a0e65f9
3
  size 5240