Training in progress, epoch 2
Browse files- model.safetensors +1 -1
- runs/Mar20_07-41-52_d04eab29c3a2/events.out.tfevents.1710920513.d04eab29c3a2.379.7 +2 -2
- runs/Mar20_08-27-43_d04eab29c3a2/events.out.tfevents.1710923265.d04eab29c3a2.379.8 +3 -0
- tmp-checkpoint-6200/config.json +48 -0
- tmp-checkpoint-6200/model.safetensors +3 -0
- tmp-checkpoint-6200/optimizer.pt +3 -0
- tmp-checkpoint-6200/preprocessor_config.json +22 -0
- tmp-checkpoint-6200/rng_state.pth +3 -0
- tmp-checkpoint-6200/scheduler.pt +3 -0
- tmp-checkpoint-6200/trainer_state.json +517 -0
- tmp-checkpoint-6200/training_args.bin +3 -0
- training_args.bin +1 -1
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 343248584
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93f0c91bb530ca1d7eeef2356d56316422ae061e540e8b55443e683f69d5f7b5
|
| 3 |
size 343248584
|
runs/Mar20_07-41-52_d04eab29c3a2/events.out.tfevents.1710920513.d04eab29c3a2.379.7
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b995f4e6a4d17fcf70dc173f78b4e5661684e8574456e1bec03e3f775c443fe7
|
| 3 |
+
size 21507
|
runs/Mar20_08-27-43_d04eab29c3a2/events.out.tfevents.1710923265.d04eab29c3a2.379.8
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7154c73bf6c8bc2de71c9e3ffbcc61f41d0a30db190fa9bc7a830dece1f5298e
|
| 3 |
+
size 6019
|
tmp-checkpoint-6200/config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-224-in21k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.0,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "Abrasions",
|
| 13 |
+
"1": "Bruises",
|
| 14 |
+
"2": "Burns",
|
| 15 |
+
"3": "Cut",
|
| 16 |
+
"4": "Diabetic Wounds",
|
| 17 |
+
"5": "Laseration",
|
| 18 |
+
"6": "Normal",
|
| 19 |
+
"7": "Pressure Wounds",
|
| 20 |
+
"8": "Surgical Wounds",
|
| 21 |
+
"9": "Venous Wounds"
|
| 22 |
+
},
|
| 23 |
+
"image_size": 224,
|
| 24 |
+
"initializer_range": 0.02,
|
| 25 |
+
"intermediate_size": 3072,
|
| 26 |
+
"label2id": {
|
| 27 |
+
"Abrasions": "0",
|
| 28 |
+
"Bruises": "1",
|
| 29 |
+
"Burns": "2",
|
| 30 |
+
"Cut": "3",
|
| 31 |
+
"Diabetic Wounds": "4",
|
| 32 |
+
"Laseration": "5",
|
| 33 |
+
"Normal": "6",
|
| 34 |
+
"Pressure Wounds": "7",
|
| 35 |
+
"Surgical Wounds": "8",
|
| 36 |
+
"Venous Wounds": "9"
|
| 37 |
+
},
|
| 38 |
+
"layer_norm_eps": 1e-12,
|
| 39 |
+
"model_type": "vit",
|
| 40 |
+
"num_attention_heads": 12,
|
| 41 |
+
"num_channels": 3,
|
| 42 |
+
"num_hidden_layers": 12,
|
| 43 |
+
"patch_size": 16,
|
| 44 |
+
"problem_type": "single_label_classification",
|
| 45 |
+
"qkv_bias": true,
|
| 46 |
+
"torch_dtype": "float32",
|
| 47 |
+
"transformers_version": "4.38.2"
|
| 48 |
+
}
|
tmp-checkpoint-6200/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc1cd4ee8f282d77bc7df90b99378c4a13f641f8bc8148de94a6230637d7ab00
|
| 3 |
+
size 343248584
|
tmp-checkpoint-6200/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f4da74c53499b252f9d5b0df838283153b1a0e0309231a1f0f05389a1b135f49
|
| 3 |
+
size 686618042
|
tmp-checkpoint-6200/preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.5,
|
| 7 |
+
0.5,
|
| 8 |
+
0.5
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.5,
|
| 13 |
+
0.5,
|
| 14 |
+
0.5
|
| 15 |
+
],
|
| 16 |
+
"resample": 2,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
tmp-checkpoint-6200/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:20ffb742a2f323b5cb46fd532016f93f52e1d2cfb8184ab08fcd3255c2a52f36
|
| 3 |
+
size 14244
|
tmp-checkpoint-6200/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77303e889270fd5ec77130e294b652df0ec5b49e051579b6c8af9176be5d6b5c
|
| 3 |
+
size 1064
|
tmp-checkpoint-6200/trainer_state.json
ADDED
|
@@ -0,0 +1,517 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9675,
|
| 3 |
+
"best_model_checkpoint": "Wound-classification/checkpoint-6000",
|
| 4 |
+
"epoch": 31.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 6200,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"grad_norm": 2.2730700969696045,
|
| 14 |
+
"learning_rate": 0.0001957070707070707,
|
| 15 |
+
"loss": 1.2288,
|
| 16 |
+
"step": 200
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_accuracy": 0.775,
|
| 21 |
+
"eval_loss": 0.7021821737289429,
|
| 22 |
+
"eval_runtime": 7.7413,
|
| 23 |
+
"eval_samples_per_second": 51.671,
|
| 24 |
+
"eval_steps_per_second": 12.918,
|
| 25 |
+
"step": 200
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 2.0,
|
| 29 |
+
"grad_norm": 0.5704953670501709,
|
| 30 |
+
"learning_rate": 0.00018939393939393942,
|
| 31 |
+
"loss": 0.6343,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 2.0,
|
| 36 |
+
"eval_accuracy": 0.7875,
|
| 37 |
+
"eval_loss": 0.6516726613044739,
|
| 38 |
+
"eval_runtime": 7.6496,
|
| 39 |
+
"eval_samples_per_second": 52.29,
|
| 40 |
+
"eval_steps_per_second": 13.073,
|
| 41 |
+
"step": 400
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 3.0,
|
| 45 |
+
"grad_norm": 6.674056053161621,
|
| 46 |
+
"learning_rate": 0.0001830808080808081,
|
| 47 |
+
"loss": 0.5022,
|
| 48 |
+
"step": 600
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 3.0,
|
| 52 |
+
"eval_accuracy": 0.73,
|
| 53 |
+
"eval_loss": 0.8923927545547485,
|
| 54 |
+
"eval_runtime": 7.8144,
|
| 55 |
+
"eval_samples_per_second": 51.187,
|
| 56 |
+
"eval_steps_per_second": 12.797,
|
| 57 |
+
"step": 600
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 4.0,
|
| 61 |
+
"grad_norm": 3.7206950187683105,
|
| 62 |
+
"learning_rate": 0.0001767676767676768,
|
| 63 |
+
"loss": 0.5071,
|
| 64 |
+
"step": 800
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 4.0,
|
| 68 |
+
"eval_accuracy": 0.8475,
|
| 69 |
+
"eval_loss": 0.46834319829940796,
|
| 70 |
+
"eval_runtime": 7.622,
|
| 71 |
+
"eval_samples_per_second": 52.48,
|
| 72 |
+
"eval_steps_per_second": 13.12,
|
| 73 |
+
"step": 800
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 5.0,
|
| 77 |
+
"grad_norm": 1.3697607517242432,
|
| 78 |
+
"learning_rate": 0.00017045454545454547,
|
| 79 |
+
"loss": 0.3484,
|
| 80 |
+
"step": 1000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 5.0,
|
| 84 |
+
"eval_accuracy": 0.8675,
|
| 85 |
+
"eval_loss": 0.44843006134033203,
|
| 86 |
+
"eval_runtime": 7.8641,
|
| 87 |
+
"eval_samples_per_second": 50.864,
|
| 88 |
+
"eval_steps_per_second": 12.716,
|
| 89 |
+
"step": 1000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 6.0,
|
| 93 |
+
"grad_norm": 4.369117259979248,
|
| 94 |
+
"learning_rate": 0.00016414141414141414,
|
| 95 |
+
"loss": 0.3452,
|
| 96 |
+
"step": 1200
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 6.0,
|
| 100 |
+
"eval_accuracy": 0.8925,
|
| 101 |
+
"eval_loss": 0.37597933411598206,
|
| 102 |
+
"eval_runtime": 7.6457,
|
| 103 |
+
"eval_samples_per_second": 52.317,
|
| 104 |
+
"eval_steps_per_second": 13.079,
|
| 105 |
+
"step": 1200
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 7.0,
|
| 109 |
+
"grad_norm": 7.87866735458374,
|
| 110 |
+
"learning_rate": 0.00015782828282828284,
|
| 111 |
+
"loss": 0.2461,
|
| 112 |
+
"step": 1400
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 7.0,
|
| 116 |
+
"eval_accuracy": 0.83,
|
| 117 |
+
"eval_loss": 0.5486177206039429,
|
| 118 |
+
"eval_runtime": 7.9064,
|
| 119 |
+
"eval_samples_per_second": 50.592,
|
| 120 |
+
"eval_steps_per_second": 12.648,
|
| 121 |
+
"step": 1400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 8.0,
|
| 125 |
+
"grad_norm": 9.048334121704102,
|
| 126 |
+
"learning_rate": 0.00015151515151515152,
|
| 127 |
+
"loss": 0.2389,
|
| 128 |
+
"step": 1600
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.0,
|
| 132 |
+
"eval_accuracy": 0.895,
|
| 133 |
+
"eval_loss": 0.40351569652557373,
|
| 134 |
+
"eval_runtime": 7.8304,
|
| 135 |
+
"eval_samples_per_second": 51.083,
|
| 136 |
+
"eval_steps_per_second": 12.771,
|
| 137 |
+
"step": 1600
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 9.0,
|
| 141 |
+
"grad_norm": 2.520663022994995,
|
| 142 |
+
"learning_rate": 0.00014520202020202022,
|
| 143 |
+
"loss": 0.2272,
|
| 144 |
+
"step": 1800
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 9.0,
|
| 148 |
+
"eval_accuracy": 0.8875,
|
| 149 |
+
"eval_loss": 0.4761699438095093,
|
| 150 |
+
"eval_runtime": 7.8414,
|
| 151 |
+
"eval_samples_per_second": 51.011,
|
| 152 |
+
"eval_steps_per_second": 12.753,
|
| 153 |
+
"step": 1800
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 10.0,
|
| 157 |
+
"grad_norm": 0.026233581826090813,
|
| 158 |
+
"learning_rate": 0.0001388888888888889,
|
| 159 |
+
"loss": 0.2008,
|
| 160 |
+
"step": 2000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 10.0,
|
| 164 |
+
"eval_accuracy": 0.905,
|
| 165 |
+
"eval_loss": 0.35686829686164856,
|
| 166 |
+
"eval_runtime": 7.6276,
|
| 167 |
+
"eval_samples_per_second": 52.441,
|
| 168 |
+
"eval_steps_per_second": 13.11,
|
| 169 |
+
"step": 2000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 11.0,
|
| 173 |
+
"grad_norm": 0.08601810783147812,
|
| 174 |
+
"learning_rate": 0.00013257575757575756,
|
| 175 |
+
"loss": 0.2173,
|
| 176 |
+
"step": 2200
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 11.0,
|
| 180 |
+
"eval_accuracy": 0.9075,
|
| 181 |
+
"eval_loss": 0.41256073117256165,
|
| 182 |
+
"eval_runtime": 7.8809,
|
| 183 |
+
"eval_samples_per_second": 50.756,
|
| 184 |
+
"eval_steps_per_second": 12.689,
|
| 185 |
+
"step": 2200
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 12.0,
|
| 189 |
+
"grad_norm": 0.09381477534770966,
|
| 190 |
+
"learning_rate": 0.00012626262626262626,
|
| 191 |
+
"loss": 0.182,
|
| 192 |
+
"step": 2400
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 12.0,
|
| 196 |
+
"eval_accuracy": 0.9125,
|
| 197 |
+
"eval_loss": 0.3593887984752655,
|
| 198 |
+
"eval_runtime": 7.8239,
|
| 199 |
+
"eval_samples_per_second": 51.125,
|
| 200 |
+
"eval_steps_per_second": 12.781,
|
| 201 |
+
"step": 2400
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 13.0,
|
| 205 |
+
"grad_norm": 23.23141098022461,
|
| 206 |
+
"learning_rate": 0.00011994949494949495,
|
| 207 |
+
"loss": 0.1622,
|
| 208 |
+
"step": 2600
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 13.0,
|
| 212 |
+
"eval_accuracy": 0.89,
|
| 213 |
+
"eval_loss": 0.3750649392604828,
|
| 214 |
+
"eval_runtime": 7.7578,
|
| 215 |
+
"eval_samples_per_second": 51.561,
|
| 216 |
+
"eval_steps_per_second": 12.89,
|
| 217 |
+
"step": 2600
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 14.0,
|
| 221 |
+
"grad_norm": 0.19918544590473175,
|
| 222 |
+
"learning_rate": 0.00011363636363636365,
|
| 223 |
+
"loss": 0.1573,
|
| 224 |
+
"step": 2800
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 14.0,
|
| 228 |
+
"eval_accuracy": 0.92,
|
| 229 |
+
"eval_loss": 0.33412453532218933,
|
| 230 |
+
"eval_runtime": 7.6756,
|
| 231 |
+
"eval_samples_per_second": 52.113,
|
| 232 |
+
"eval_steps_per_second": 13.028,
|
| 233 |
+
"step": 2800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 15.0,
|
| 237 |
+
"grad_norm": 5.899987697601318,
|
| 238 |
+
"learning_rate": 0.00010732323232323234,
|
| 239 |
+
"loss": 0.148,
|
| 240 |
+
"step": 3000
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 15.0,
|
| 244 |
+
"eval_accuracy": 0.935,
|
| 245 |
+
"eval_loss": 0.2885863184928894,
|
| 246 |
+
"eval_runtime": 7.6591,
|
| 247 |
+
"eval_samples_per_second": 52.226,
|
| 248 |
+
"eval_steps_per_second": 13.056,
|
| 249 |
+
"step": 3000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 16.0,
|
| 253 |
+
"grad_norm": 3.5471909046173096,
|
| 254 |
+
"learning_rate": 0.00010101010101010102,
|
| 255 |
+
"loss": 0.1234,
|
| 256 |
+
"step": 3200
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 16.0,
|
| 260 |
+
"eval_accuracy": 0.9075,
|
| 261 |
+
"eval_loss": 0.3466532230377197,
|
| 262 |
+
"eval_runtime": 7.8092,
|
| 263 |
+
"eval_samples_per_second": 51.222,
|
| 264 |
+
"eval_steps_per_second": 12.805,
|
| 265 |
+
"step": 3200
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 17.0,
|
| 269 |
+
"grad_norm": 9.151985168457031,
|
| 270 |
+
"learning_rate": 9.469696969696971e-05,
|
| 271 |
+
"loss": 0.1254,
|
| 272 |
+
"step": 3400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 17.0,
|
| 276 |
+
"eval_accuracy": 0.925,
|
| 277 |
+
"eval_loss": 0.33074432611465454,
|
| 278 |
+
"eval_runtime": 7.9693,
|
| 279 |
+
"eval_samples_per_second": 50.193,
|
| 280 |
+
"eval_steps_per_second": 12.548,
|
| 281 |
+
"step": 3400
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 18.0,
|
| 285 |
+
"grad_norm": 0.017998667433857918,
|
| 286 |
+
"learning_rate": 8.83838383838384e-05,
|
| 287 |
+
"loss": 0.1104,
|
| 288 |
+
"step": 3600
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 18.0,
|
| 292 |
+
"eval_accuracy": 0.9375,
|
| 293 |
+
"eval_loss": 0.2563402056694031,
|
| 294 |
+
"eval_runtime": 7.8711,
|
| 295 |
+
"eval_samples_per_second": 50.819,
|
| 296 |
+
"eval_steps_per_second": 12.705,
|
| 297 |
+
"step": 3600
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 19.0,
|
| 301 |
+
"grad_norm": 0.07069282978773117,
|
| 302 |
+
"learning_rate": 8.207070707070707e-05,
|
| 303 |
+
"loss": 0.1017,
|
| 304 |
+
"step": 3800
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 19.0,
|
| 308 |
+
"eval_accuracy": 0.9525,
|
| 309 |
+
"eval_loss": 0.23432372510433197,
|
| 310 |
+
"eval_runtime": 8.1296,
|
| 311 |
+
"eval_samples_per_second": 49.203,
|
| 312 |
+
"eval_steps_per_second": 12.301,
|
| 313 |
+
"step": 3800
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 20.0,
|
| 317 |
+
"grad_norm": 0.00819435901939869,
|
| 318 |
+
"learning_rate": 7.575757575757576e-05,
|
| 319 |
+
"loss": 0.0912,
|
| 320 |
+
"step": 4000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 20.0,
|
| 324 |
+
"eval_accuracy": 0.925,
|
| 325 |
+
"eval_loss": 0.2641538679599762,
|
| 326 |
+
"eval_runtime": 7.8726,
|
| 327 |
+
"eval_samples_per_second": 50.809,
|
| 328 |
+
"eval_steps_per_second": 12.702,
|
| 329 |
+
"step": 4000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 21.0,
|
| 333 |
+
"grad_norm": 0.1772022545337677,
|
| 334 |
+
"learning_rate": 6.944444444444444e-05,
|
| 335 |
+
"loss": 0.093,
|
| 336 |
+
"step": 4200
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 21.0,
|
| 340 |
+
"eval_accuracy": 0.9375,
|
| 341 |
+
"eval_loss": 0.2690765857696533,
|
| 342 |
+
"eval_runtime": 7.671,
|
| 343 |
+
"eval_samples_per_second": 52.144,
|
| 344 |
+
"eval_steps_per_second": 13.036,
|
| 345 |
+
"step": 4200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 22.0,
|
| 349 |
+
"grad_norm": 6.940694332122803,
|
| 350 |
+
"learning_rate": 6.313131313131313e-05,
|
| 351 |
+
"loss": 0.1264,
|
| 352 |
+
"step": 4400
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 22.0,
|
| 356 |
+
"eval_accuracy": 0.9625,
|
| 357 |
+
"eval_loss": 0.18011243641376495,
|
| 358 |
+
"eval_runtime": 8.061,
|
| 359 |
+
"eval_samples_per_second": 49.622,
|
| 360 |
+
"eval_steps_per_second": 12.405,
|
| 361 |
+
"step": 4400
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 23.0,
|
| 365 |
+
"grad_norm": 0.007489996496587992,
|
| 366 |
+
"learning_rate": 5.6818181818181825e-05,
|
| 367 |
+
"loss": 0.0951,
|
| 368 |
+
"step": 4600
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 23.0,
|
| 372 |
+
"eval_accuracy": 0.93,
|
| 373 |
+
"eval_loss": 0.3289620280265808,
|
| 374 |
+
"eval_runtime": 7.6993,
|
| 375 |
+
"eval_samples_per_second": 51.953,
|
| 376 |
+
"eval_steps_per_second": 12.988,
|
| 377 |
+
"step": 4600
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 24.0,
|
| 381 |
+
"grad_norm": 0.008240079507231712,
|
| 382 |
+
"learning_rate": 5.050505050505051e-05,
|
| 383 |
+
"loss": 0.0793,
|
| 384 |
+
"step": 4800
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 24.0,
|
| 388 |
+
"eval_accuracy": 0.935,
|
| 389 |
+
"eval_loss": 0.30677276849746704,
|
| 390 |
+
"eval_runtime": 8.0535,
|
| 391 |
+
"eval_samples_per_second": 49.668,
|
| 392 |
+
"eval_steps_per_second": 12.417,
|
| 393 |
+
"step": 4800
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 25.0,
|
| 397 |
+
"grad_norm": 0.06566061079502106,
|
| 398 |
+
"learning_rate": 4.41919191919192e-05,
|
| 399 |
+
"loss": 0.0489,
|
| 400 |
+
"step": 5000
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 25.0,
|
| 404 |
+
"eval_accuracy": 0.9575,
|
| 405 |
+
"eval_loss": 0.2502322494983673,
|
| 406 |
+
"eval_runtime": 7.9899,
|
| 407 |
+
"eval_samples_per_second": 50.063,
|
| 408 |
+
"eval_steps_per_second": 12.516,
|
| 409 |
+
"step": 5000
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 26.0,
|
| 413 |
+
"grad_norm": 0.006019990891218185,
|
| 414 |
+
"learning_rate": 3.787878787878788e-05,
|
| 415 |
+
"loss": 0.0824,
|
| 416 |
+
"step": 5200
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 26.0,
|
| 420 |
+
"eval_accuracy": 0.9425,
|
| 421 |
+
"eval_loss": 0.27752983570098877,
|
| 422 |
+
"eval_runtime": 7.7875,
|
| 423 |
+
"eval_samples_per_second": 51.364,
|
| 424 |
+
"eval_steps_per_second": 12.841,
|
| 425 |
+
"step": 5200
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 27.0,
|
| 429 |
+
"grad_norm": 0.006204684264957905,
|
| 430 |
+
"learning_rate": 3.1565656565656566e-05,
|
| 431 |
+
"loss": 0.0784,
|
| 432 |
+
"step": 5400
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 27.0,
|
| 436 |
+
"eval_accuracy": 0.9525,
|
| 437 |
+
"eval_loss": 0.21709877252578735,
|
| 438 |
+
"eval_runtime": 7.8619,
|
| 439 |
+
"eval_samples_per_second": 50.878,
|
| 440 |
+
"eval_steps_per_second": 12.72,
|
| 441 |
+
"step": 5400
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 28.0,
|
| 445 |
+
"grad_norm": 0.436471164226532,
|
| 446 |
+
"learning_rate": 2.5252525252525256e-05,
|
| 447 |
+
"loss": 0.0463,
|
| 448 |
+
"step": 5600
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 28.0,
|
| 452 |
+
"eval_accuracy": 0.96,
|
| 453 |
+
"eval_loss": 0.22668342292308807,
|
| 454 |
+
"eval_runtime": 7.8277,
|
| 455 |
+
"eval_samples_per_second": 51.101,
|
| 456 |
+
"eval_steps_per_second": 12.775,
|
| 457 |
+
"step": 5600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 29.0,
|
| 461 |
+
"grad_norm": 0.006457165349274874,
|
| 462 |
+
"learning_rate": 1.893939393939394e-05,
|
| 463 |
+
"loss": 0.0553,
|
| 464 |
+
"step": 5800
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 29.0,
|
| 468 |
+
"eval_accuracy": 0.96,
|
| 469 |
+
"eval_loss": 0.20785684883594513,
|
| 470 |
+
"eval_runtime": 7.5766,
|
| 471 |
+
"eval_samples_per_second": 52.794,
|
| 472 |
+
"eval_steps_per_second": 13.199,
|
| 473 |
+
"step": 5800
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 30.0,
|
| 477 |
+
"grad_norm": 5.3711652755737305,
|
| 478 |
+
"learning_rate": 1.2626262626262628e-05,
|
| 479 |
+
"loss": 0.0247,
|
| 480 |
+
"step": 6000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 30.0,
|
| 484 |
+
"eval_accuracy": 0.9675,
|
| 485 |
+
"eval_loss": 0.1625244915485382,
|
| 486 |
+
"eval_runtime": 7.7936,
|
| 487 |
+
"eval_samples_per_second": 51.324,
|
| 488 |
+
"eval_steps_per_second": 12.831,
|
| 489 |
+
"step": 6000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 31.0,
|
| 493 |
+
"grad_norm": 0.718065619468689,
|
| 494 |
+
"learning_rate": 6.313131313131314e-06,
|
| 495 |
+
"loss": 0.0356,
|
| 496 |
+
"step": 6200
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 31.0,
|
| 500 |
+
"eval_accuracy": 0.9575,
|
| 501 |
+
"eval_loss": 0.2557324469089508,
|
| 502 |
+
"eval_runtime": 7.6107,
|
| 503 |
+
"eval_samples_per_second": 52.557,
|
| 504 |
+
"eval_steps_per_second": 13.139,
|
| 505 |
+
"step": 6200
|
| 506 |
+
}
|
| 507 |
+
],
|
| 508 |
+
"logging_steps": 500,
|
| 509 |
+
"max_steps": 6400,
|
| 510 |
+
"num_input_tokens_seen": 0,
|
| 511 |
+
"num_train_epochs": 32,
|
| 512 |
+
"save_steps": 500,
|
| 513 |
+
"total_flos": 3.843878276844749e+18,
|
| 514 |
+
"train_batch_size": 4,
|
| 515 |
+
"trial_name": null,
|
| 516 |
+
"trial_params": null
|
| 517 |
+
}
|
tmp-checkpoint-6200/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:258d4988c92e75f79fdad65f22db228fd4e6c6404660958517836a7f03f590ea
|
| 3 |
+
size 4920
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4920
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6b9e88cfbfb67548abcb085b4dfc5d0ead8b356d7fd3a43b4ef00f22729a9fe
|
| 3 |
size 4920
|