Training in progress, epoch 1, checkpoint
Browse files
checkpoint-3743/generation_config.json
CHANGED
|
@@ -4,8 +4,9 @@
|
|
| 4 |
"early_stopping": true,
|
| 5 |
"eos_token_id": 2,
|
| 6 |
"forced_eos_token_id": 2,
|
|
|
|
| 7 |
"no_repeat_ngram_size": 3,
|
| 8 |
-
"num_beams":
|
| 9 |
"pad_token_id": 1,
|
| 10 |
"transformers_version": "4.42.3"
|
| 11 |
}
|
|
|
|
| 4 |
"early_stopping": true,
|
| 5 |
"eos_token_id": 2,
|
| 6 |
"forced_eos_token_id": 2,
|
| 7 |
+
"max_new_tokens": 100,
|
| 8 |
"no_repeat_ngram_size": 3,
|
| 9 |
+
"num_beams": 3,
|
| 10 |
"pad_token_id": 1,
|
| 11 |
"transformers_version": "4.42.3"
|
| 12 |
}
|
checkpoint-3743/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 557116312
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b824504b5a81b36d7b259896b5d6b8d4887d1d40457a03a884e8fa64abb7e6d
|
| 3 |
size 557116312
|
checkpoint-3743/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1113991930
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71578d1e9ef3bd5cce555d097c5822f4c19c6fb19489369d66e541fac12361a1
|
| 3 |
size 1113991930
|
checkpoint-3743/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ff10cc3496c08e1106e5e897405ff4505ff584fdbc5ddf9c1ac1e5aa55049cd1
|
| 3 |
size 14244
|
checkpoint-3743/trainer_state.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
"best_model_checkpoint": "/content/drive/MyDrive/Arabart/checkpoint-3743",
|
| 4 |
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
|
@@ -10,62 +10,62 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.1335826876836762,
|
| 13 |
-
"grad_norm": 5.
|
| 14 |
"learning_rate": 3.7102997922232118e-06,
|
| 15 |
-
"loss":
|
| 16 |
"step": 500
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.2671653753673524,
|
| 20 |
-
"grad_norm":
|
| 21 |
"learning_rate": 7.4205995844464235e-06,
|
| 22 |
-
"loss": 3.
|
| 23 |
"step": 1000
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.4007480630510286,
|
| 27 |
-
"grad_norm": 3.
|
| 28 |
"learning_rate": 1.1130899376669635e-05,
|
| 29 |
-
"loss": 3.
|
| 30 |
"step": 1500
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.5343307507347048,
|
| 34 |
-
"grad_norm": 4.
|
| 35 |
"learning_rate": 1.4841199168892847e-05,
|
| 36 |
-
"loss":
|
| 37 |
"step": 2000
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.667913438418381,
|
| 41 |
-
"grad_norm":
|
| 42 |
"learning_rate": 1.855149896111606e-05,
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 2500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.8014961261020572,
|
| 48 |
-
"grad_norm": 3.
|
| 49 |
"learning_rate": 2.226179875333927e-05,
|
| 50 |
-
"loss": 2.
|
| 51 |
"step": 3000
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.9350788137857333,
|
| 55 |
-
"grad_norm": 3.
|
| 56 |
"learning_rate": 2.5972098545562486e-05,
|
| 57 |
-
"loss": 2.
|
| 58 |
"step": 3500
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 1.0,
|
| 62 |
-
"eval_loss": 2.
|
| 63 |
-
"eval_rouge1":
|
| 64 |
-
"eval_rouge2":
|
| 65 |
-
"eval_rougeL":
|
| 66 |
-
"eval_runtime":
|
| 67 |
-
"eval_samples_per_second":
|
| 68 |
-
"eval_steps_per_second": 0.
|
| 69 |
"step": 3743
|
| 70 |
}
|
| 71 |
],
|
|
@@ -86,7 +86,7 @@
|
|
| 86 |
"attributes": {}
|
| 87 |
}
|
| 88 |
},
|
| 89 |
-
"total_flos": 2.
|
| 90 |
"train_batch_size": 10,
|
| 91 |
"trial_name": null,
|
| 92 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 25.9968,
|
| 3 |
"best_model_checkpoint": "/content/drive/MyDrive/Arabart/checkpoint-3743",
|
| 4 |
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.1335826876836762,
|
| 13 |
+
"grad_norm": 5.345764636993408,
|
| 14 |
"learning_rate": 3.7102997922232118e-06,
|
| 15 |
+
"loss": 3.9988,
|
| 16 |
"step": 500
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.2671653753673524,
|
| 20 |
+
"grad_norm": 5.3753743171691895,
|
| 21 |
"learning_rate": 7.4205995844464235e-06,
|
| 22 |
+
"loss": 3.2671,
|
| 23 |
"step": 1000
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.4007480630510286,
|
| 27 |
+
"grad_norm": 3.7295212745666504,
|
| 28 |
"learning_rate": 1.1130899376669635e-05,
|
| 29 |
+
"loss": 3.1006,
|
| 30 |
"step": 1500
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.5343307507347048,
|
| 34 |
+
"grad_norm": 4.027897834777832,
|
| 35 |
"learning_rate": 1.4841199168892847e-05,
|
| 36 |
+
"loss": 2.9767,
|
| 37 |
"step": 2000
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.667913438418381,
|
| 41 |
+
"grad_norm": 3.9697580337524414,
|
| 42 |
"learning_rate": 1.855149896111606e-05,
|
| 43 |
+
"loss": 2.8914,
|
| 44 |
"step": 2500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.8014961261020572,
|
| 48 |
+
"grad_norm": 3.5195724964141846,
|
| 49 |
"learning_rate": 2.226179875333927e-05,
|
| 50 |
+
"loss": 2.8273,
|
| 51 |
"step": 3000
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.9350788137857333,
|
| 55 |
+
"grad_norm": 3.447735071182251,
|
| 56 |
"learning_rate": 2.5972098545562486e-05,
|
| 57 |
+
"loss": 2.7964,
|
| 58 |
"step": 3500
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 1.0,
|
| 62 |
+
"eval_loss": 2.2515146732330322,
|
| 63 |
+
"eval_rouge1": 25.9968,
|
| 64 |
+
"eval_rouge2": 10.7383,
|
| 65 |
+
"eval_rougeL": 22.8305,
|
| 66 |
+
"eval_runtime": 801.3334,
|
| 67 |
+
"eval_samples_per_second": 5.851,
|
| 68 |
+
"eval_steps_per_second": 0.585,
|
| 69 |
"step": 3743
|
| 70 |
}
|
| 71 |
],
|
|
|
|
| 86 |
"attributes": {}
|
| 87 |
}
|
| 88 |
},
|
| 89 |
+
"total_flos": 2.191707783032832e+16,
|
| 90 |
"train_batch_size": 10,
|
| 91 |
"trial_name": null,
|
| 92 |
"trial_params": null
|
checkpoint-3743/training_args.bin
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 sha256:0af844ab23c78977e15af3b0c2d9e2fc6623c1ae2c85b79f47e563f42dc7371b
|
| 3 |
+
size 7032
|