Upload cross-lingual sentiment model
Browse files- model.safetensors +1 -1
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +90 -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 1112205008
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb45b67ccbe2dba6dc275e4943154d50c84b6de13020fd0ad7959ec268ac9fdf
|
| 3 |
size 1112205008
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:271d4cd0e097165570a879ddd0600ed11542ccb32f1d6dd469d93470167544f9
|
| 3 |
+
size 2224524427
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:41cfd668bbf9b1ccfec9d84d993e978b93d745f8f881e6aae24dac508151d8e7
|
| 3 |
+
size 14455
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c3989decec577ddb21c6af650a78290b7e5a08092908de2f76a6b266620e2ea
|
| 3 |
+
size 1465
|
trainer_state.json
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 500,
|
| 3 |
+
"best_metric": 1.0,
|
| 4 |
+
"best_model_checkpoint": "experiments/combined/checkpoint-500",
|
| 5 |
+
"epoch": 0.9652509652509652,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.19305019305019305,
|
| 14 |
+
"grad_norm": 2.48384165763855,
|
| 15 |
+
"learning_rate": 1.2692307692307693e-05,
|
| 16 |
+
"loss": 0.4237,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3861003861003861,
|
| 21 |
+
"grad_norm": 6.34004020690918,
|
| 22 |
+
"learning_rate": 1.938483547925608e-05,
|
| 23 |
+
"loss": 0.2115,
|
| 24 |
+
"step": 200
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.5791505791505791,
|
| 28 |
+
"grad_norm": 0.6042028665542603,
|
| 29 |
+
"learning_rate": 1.795422031473534e-05,
|
| 30 |
+
"loss": 0.1104,
|
| 31 |
+
"step": 300
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.7722007722007722,
|
| 35 |
+
"grad_norm": 0.008698650635778904,
|
| 36 |
+
"learning_rate": 1.6523605150214594e-05,
|
| 37 |
+
"loss": 0.0005,
|
| 38 |
+
"step": 400
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.9652509652509652,
|
| 42 |
+
"grad_norm": 20.775955200195312,
|
| 43 |
+
"learning_rate": 1.509298998569385e-05,
|
| 44 |
+
"loss": 0.0041,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9652509652509652,
|
| 49 |
+
"eval_accuracy": 1.0,
|
| 50 |
+
"eval_f1": 1.0,
|
| 51 |
+
"eval_loss": 4.6818098780931905e-05,
|
| 52 |
+
"eval_precision": 1.0,
|
| 53 |
+
"eval_recall": 1.0,
|
| 54 |
+
"eval_runtime": 52.5745,
|
| 55 |
+
"eval_samples_per_second": 8.749,
|
| 56 |
+
"eval_steps_per_second": 0.285,
|
| 57 |
+
"step": 500
|
| 58 |
+
}
|
| 59 |
+
],
|
| 60 |
+
"logging_steps": 100,
|
| 61 |
+
"max_steps": 1554,
|
| 62 |
+
"num_input_tokens_seen": 0,
|
| 63 |
+
"num_train_epochs": 3,
|
| 64 |
+
"save_steps": 500,
|
| 65 |
+
"stateful_callbacks": {
|
| 66 |
+
"EarlyStoppingCallback": {
|
| 67 |
+
"args": {
|
| 68 |
+
"early_stopping_patience": 3,
|
| 69 |
+
"early_stopping_threshold": 0.0
|
| 70 |
+
},
|
| 71 |
+
"attributes": {
|
| 72 |
+
"early_stopping_patience_counter": 0
|
| 73 |
+
}
|
| 74 |
+
},
|
| 75 |
+
"TrainerControl": {
|
| 76 |
+
"args": {
|
| 77 |
+
"should_epoch_stop": false,
|
| 78 |
+
"should_evaluate": false,
|
| 79 |
+
"should_log": false,
|
| 80 |
+
"should_save": true,
|
| 81 |
+
"should_training_stop": false
|
| 82 |
+
},
|
| 83 |
+
"attributes": {}
|
| 84 |
+
}
|
| 85 |
+
},
|
| 86 |
+
"total_flos": 263111055360000.0,
|
| 87 |
+
"train_batch_size": 8,
|
| 88 |
+
"trial_name": null,
|
| 89 |
+
"trial_params": null
|
| 90 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5777
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d87db1a9830ea23deae1ab503d8bfe746de6b0751e5561d09a6dfeaf53f7d5a4
|
| 3 |
size 5777
|