Commit
•
b64856f
1
Parent(s):
6d17ebb
Training in progress, step 1000
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- model.safetensors +1 -1
- run-0/checkpoint-1000/model.safetensors +1 -1
- run-0/checkpoint-1000/optimizer.pt +1 -1
- run-0/checkpoint-1000/scheduler.pt +1 -1
- run-0/checkpoint-1000/trainer_state.json +27 -27
- run-0/checkpoint-1000/training_args.bin +1 -1
- run-0/checkpoint-1500/trainer_state.json +35 -35
- run-0/checkpoint-2000/trainer_state.json +48 -48
- run-0/checkpoint-500/model.safetensors +1 -1
- run-0/checkpoint-500/optimizer.pt +1 -1
- run-0/checkpoint-500/scheduler.pt +1 -1
- run-0/checkpoint-500/trainer_state.json +14 -14
- run-0/checkpoint-500/training_args.bin +1 -1
- run-7/checkpoint-2500/config.json +331 -0
- run-7/checkpoint-2500/model.safetensors +3 -0
- run-7/checkpoint-2500/optimizer.pt +3 -0
- run-7/checkpoint-2500/rng_state.pth +3 -0
- run-7/checkpoint-2500/scheduler.pt +3 -0
- run-7/checkpoint-2500/special_tokens_map.json +7 -0
- run-7/checkpoint-2500/tokenizer.json +0 -0
- run-7/checkpoint-2500/tokenizer_config.json +55 -0
- run-7/checkpoint-2500/trainer_state.json +135 -0
- run-7/checkpoint-2500/training_args.bin +3 -0
- run-7/checkpoint-2500/vocab.txt +0 -0
- run-7/checkpoint-2862/config.json +331 -0
- run-7/checkpoint-2862/model.safetensors +3 -0
- run-7/checkpoint-2862/optimizer.pt +3 -0
- run-7/checkpoint-2862/rng_state.pth +3 -0
- run-7/checkpoint-2862/scheduler.pt +3 -0
- run-7/checkpoint-2862/special_tokens_map.json +7 -0
- run-7/checkpoint-2862/tokenizer.json +0 -0
- run-7/checkpoint-2862/tokenizer_config.json +55 -0
- run-7/checkpoint-2862/trainer_state.json +144 -0
- run-7/checkpoint-2862/training_args.bin +3 -0
- run-7/checkpoint-2862/vocab.txt +0 -0
- run-8/checkpoint-1000/config.json +331 -0
- run-8/checkpoint-1000/model.safetensors +3 -0
- run-8/checkpoint-1000/optimizer.pt +3 -0
- run-8/checkpoint-1000/rng_state.pth +3 -0
- run-8/checkpoint-1000/scheduler.pt +3 -0
- run-8/checkpoint-1000/special_tokens_map.json +7 -0
- run-8/checkpoint-1000/tokenizer.json +0 -0
- run-8/checkpoint-1000/tokenizer_config.json +55 -0
- run-8/checkpoint-1000/trainer_state.json +78 -0
- run-8/checkpoint-1000/training_args.bin +3 -0
- run-8/checkpoint-1000/vocab.txt +0 -0
- run-8/checkpoint-500/config.json +331 -0
- run-8/checkpoint-500/model.safetensors +3 -0
- run-8/checkpoint-500/optimizer.pt +3 -0
- run-8/checkpoint-500/rng_state.pth +3 -0
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 268290900
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04369c6e2112f3462af4f06ced86d9ed2cbcf1bcec738f9218de55fde5f2a1b8
|
| 3 |
size 268290900
|
run-0/checkpoint-1000/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 268290900
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04369c6e2112f3462af4f06ced86d9ed2cbcf1bcec738f9218de55fde5f2a1b8
|
| 3 |
size 268290900
|
run-0/checkpoint-1000/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 536643898
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f34c742a5b7f616d81db8b28603abff514a00c8ee18aef06889fc9111b6ff98a
|
| 3 |
size 536643898
|
run-0/checkpoint-1000/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:045cf4abe065257226fe11ac1a91dfd4c5b953cb2366365ef8bb18b4cf33e734
|
| 3 |
size 1064
|
run-0/checkpoint-1000/trainer_state.json
CHANGED
|
@@ -10,50 +10,50 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
-
"eval_accuracy": 0.
|
| 14 |
-
"eval_loss": 0.
|
| 15 |
-
"eval_runtime":
|
| 16 |
-
"eval_samples_per_second":
|
| 17 |
-
"eval_steps_per_second":
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
-
"grad_norm": 0.
|
| 23 |
-
"learning_rate": 1.
|
| 24 |
-
"loss": 0.
|
| 25 |
"step": 500
|
| 26 |
},
|
| 27 |
{
|
| 28 |
"epoch": 2.0,
|
| 29 |
-
"eval_accuracy": 0.
|
| 30 |
-
"eval_loss": 0.
|
| 31 |
-
"eval_runtime":
|
| 32 |
-
"eval_samples_per_second":
|
| 33 |
-
"eval_steps_per_second": 10.
|
| 34 |
"step": 636
|
| 35 |
},
|
| 36 |
{
|
| 37 |
"epoch": 3.0,
|
| 38 |
-
"eval_accuracy": 0.
|
| 39 |
-
"eval_loss": 0.
|
| 40 |
-
"eval_runtime": 5.
|
| 41 |
-
"eval_samples_per_second":
|
| 42 |
-
"eval_steps_per_second": 11.
|
| 43 |
"step": 954
|
| 44 |
},
|
| 45 |
{
|
| 46 |
"epoch": 3.1446540880503147,
|
| 47 |
-
"grad_norm": 0.
|
| 48 |
-
"learning_rate": 1.
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 1000
|
| 51 |
}
|
| 52 |
],
|
| 53 |
"logging_steps": 500,
|
| 54 |
-
"max_steps":
|
| 55 |
"num_input_tokens_seen": 0,
|
| 56 |
-
"num_train_epochs":
|
| 57 |
"save_steps": 500,
|
| 58 |
"stateful_callbacks": {
|
| 59 |
"TrainerControl": {
|
|
@@ -67,12 +67,12 @@
|
|
| 67 |
"attributes": {}
|
| 68 |
}
|
| 69 |
},
|
| 70 |
-
"total_flos":
|
| 71 |
"train_batch_size": 48,
|
| 72 |
"trial_name": null,
|
| 73 |
"trial_params": {
|
| 74 |
-
"alpha": 0.
|
| 75 |
-
"num_train_epochs":
|
| 76 |
-
"temperature":
|
| 77 |
}
|
| 78 |
}
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.6758064516129032,
|
| 14 |
+
"eval_loss": 0.40983492136001587,
|
| 15 |
+
"eval_runtime": 5.8092,
|
| 16 |
+
"eval_samples_per_second": 533.639,
|
| 17 |
+
"eval_steps_per_second": 11.189,
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.86518794298172,
|
| 23 |
+
"learning_rate": 1.606918238993711e-05,
|
| 24 |
+
"loss": 0.645,
|
| 25 |
"step": 500
|
| 26 |
},
|
| 27 |
{
|
| 28 |
"epoch": 2.0,
|
| 29 |
+
"eval_accuracy": 0.8412903225806452,
|
| 30 |
+
"eval_loss": 0.142293781042099,
|
| 31 |
+
"eval_runtime": 6.1991,
|
| 32 |
+
"eval_samples_per_second": 500.073,
|
| 33 |
+
"eval_steps_per_second": 10.485,
|
| 34 |
"step": 636
|
| 35 |
},
|
| 36 |
{
|
| 37 |
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.8938709677419355,
|
| 39 |
+
"eval_loss": 0.07486626505851746,
|
| 40 |
+
"eval_runtime": 5.5967,
|
| 41 |
+
"eval_samples_per_second": 553.895,
|
| 42 |
+
"eval_steps_per_second": 11.614,
|
| 43 |
"step": 954
|
| 44 |
},
|
| 45 |
{
|
| 46 |
"epoch": 3.1446540880503147,
|
| 47 |
+
"grad_norm": 0.7068584561347961,
|
| 48 |
+
"learning_rate": 1.2138364779874214e-05,
|
| 49 |
+
"loss": 0.1671,
|
| 50 |
"step": 1000
|
| 51 |
}
|
| 52 |
],
|
| 53 |
"logging_steps": 500,
|
| 54 |
+
"max_steps": 2544,
|
| 55 |
"num_input_tokens_seen": 0,
|
| 56 |
+
"num_train_epochs": 8,
|
| 57 |
"save_steps": 500,
|
| 58 |
"stateful_callbacks": {
|
| 59 |
"TrainerControl": {
|
|
|
|
| 67 |
"attributes": {}
|
| 68 |
}
|
| 69 |
},
|
| 70 |
+
"total_flos": 308457471958440.0,
|
| 71 |
"train_batch_size": 48,
|
| 72 |
"trial_name": null,
|
| 73 |
"trial_params": {
|
| 74 |
+
"alpha": 0.3417365166940647,
|
| 75 |
+
"num_train_epochs": 8,
|
| 76 |
+
"temperature": 2
|
| 77 |
}
|
| 78 |
}
|
run-0/checkpoint-1000/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a795e38fc17d23ea7e491b0b97849dc60c777dbce157fcfdc837d87b44f51896
|
| 3 |
size 5176
|
run-0/checkpoint-1500/trainer_state.json
CHANGED
|
@@ -10,66 +10,66 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
-
"eval_accuracy": 0.
|
| 14 |
-
"eval_loss": 0.
|
| 15 |
-
"eval_runtime":
|
| 16 |
-
"eval_samples_per_second":
|
| 17 |
-
"eval_steps_per_second":
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
-
"grad_norm": 0.
|
| 23 |
-
"learning_rate": 1.
|
| 24 |
-
"loss": 0.
|
| 25 |
"step": 500
|
| 26 |
},
|
| 27 |
{
|
| 28 |
"epoch": 2.0,
|
| 29 |
-
"eval_accuracy": 0.
|
| 30 |
-
"eval_loss": 0.
|
| 31 |
-
"eval_runtime":
|
| 32 |
-
"eval_samples_per_second":
|
| 33 |
-
"eval_steps_per_second": 10.
|
| 34 |
"step": 636
|
| 35 |
},
|
| 36 |
{
|
| 37 |
"epoch": 3.0,
|
| 38 |
-
"eval_accuracy": 0.
|
| 39 |
-
"eval_loss": 0.
|
| 40 |
-
"eval_runtime": 5.
|
| 41 |
-
"eval_samples_per_second":
|
| 42 |
-
"eval_steps_per_second": 11.
|
| 43 |
"step": 954
|
| 44 |
},
|
| 45 |
{
|
| 46 |
"epoch": 3.1446540880503147,
|
| 47 |
-
"grad_norm": 0.
|
| 48 |
-
"learning_rate": 1.
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 1000
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 4.0,
|
| 54 |
-
"eval_accuracy": 0.
|
| 55 |
-
"eval_loss": 0.
|
| 56 |
-
"eval_runtime": 5.
|
| 57 |
-
"eval_samples_per_second":
|
| 58 |
-
"eval_steps_per_second": 11.
|
| 59 |
"step": 1272
|
| 60 |
},
|
| 61 |
{
|
| 62 |
"epoch": 4.716981132075472,
|
| 63 |
-
"grad_norm": 0.
|
| 64 |
-
"learning_rate":
|
| 65 |
-
"loss": 0.
|
| 66 |
"step": 1500
|
| 67 |
}
|
| 68 |
],
|
| 69 |
"logging_steps": 500,
|
| 70 |
-
"max_steps":
|
| 71 |
"num_input_tokens_seen": 0,
|
| 72 |
-
"num_train_epochs":
|
| 73 |
"save_steps": 500,
|
| 74 |
"stateful_callbacks": {
|
| 75 |
"TrainerControl": {
|
|
@@ -83,12 +83,12 @@
|
|
| 83 |
"attributes": {}
|
| 84 |
}
|
| 85 |
},
|
| 86 |
-
"total_flos":
|
| 87 |
"train_batch_size": 48,
|
| 88 |
"trial_name": null,
|
| 89 |
"trial_params": {
|
| 90 |
-
"alpha": 0.
|
| 91 |
-
"num_train_epochs":
|
| 92 |
-
"temperature":
|
| 93 |
}
|
| 94 |
}
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.6758064516129032,
|
| 14 |
+
"eval_loss": 0.40983492136001587,
|
| 15 |
+
"eval_runtime": 5.8092,
|
| 16 |
+
"eval_samples_per_second": 533.639,
|
| 17 |
+
"eval_steps_per_second": 11.189,
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.86518794298172,
|
| 23 |
+
"learning_rate": 1.606918238993711e-05,
|
| 24 |
+
"loss": 0.645,
|
| 25 |
"step": 500
|
| 26 |
},
|
| 27 |
{
|
| 28 |
"epoch": 2.0,
|
| 29 |
+
"eval_accuracy": 0.8412903225806452,
|
| 30 |
+
"eval_loss": 0.142293781042099,
|
| 31 |
+
"eval_runtime": 6.1991,
|
| 32 |
+
"eval_samples_per_second": 500.073,
|
| 33 |
+
"eval_steps_per_second": 10.485,
|
| 34 |
"step": 636
|
| 35 |
},
|
| 36 |
{
|
| 37 |
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.8938709677419355,
|
| 39 |
+
"eval_loss": 0.07486626505851746,
|
| 40 |
+
"eval_runtime": 5.5967,
|
| 41 |
+
"eval_samples_per_second": 553.895,
|
| 42 |
+
"eval_steps_per_second": 11.614,
|
| 43 |
"step": 954
|
| 44 |
},
|
| 45 |
{
|
| 46 |
"epoch": 3.1446540880503147,
|
| 47 |
+
"grad_norm": 0.7068584561347961,
|
| 48 |
+
"learning_rate": 1.2138364779874214e-05,
|
| 49 |
+
"loss": 0.1671,
|
| 50 |
"step": 1000
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 4.0,
|
| 54 |
+
"eval_accuracy": 0.9141935483870968,
|
| 55 |
+
"eval_loss": 0.053671687841415405,
|
| 56 |
+
"eval_runtime": 5.7734,
|
| 57 |
+
"eval_samples_per_second": 536.949,
|
| 58 |
+
"eval_steps_per_second": 11.259,
|
| 59 |
"step": 1272
|
| 60 |
},
|
| 61 |
{
|
| 62 |
"epoch": 4.716981132075472,
|
| 63 |
+
"grad_norm": 0.464497447013855,
|
| 64 |
+
"learning_rate": 8.207547169811321e-06,
|
| 65 |
+
"loss": 0.0849,
|
| 66 |
"step": 1500
|
| 67 |
}
|
| 68 |
],
|
| 69 |
"logging_steps": 500,
|
| 70 |
+
"max_steps": 2544,
|
| 71 |
"num_input_tokens_seen": 0,
|
| 72 |
+
"num_train_epochs": 8,
|
| 73 |
"save_steps": 500,
|
| 74 |
"stateful_callbacks": {
|
| 75 |
"TrainerControl": {
|
|
|
|
| 83 |
"attributes": {}
|
| 84 |
}
|
| 85 |
},
|
| 86 |
+
"total_flos": 438885076748160.0,
|
| 87 |
"train_batch_size": 48,
|
| 88 |
"trial_name": null,
|
| 89 |
"trial_params": {
|
| 90 |
+
"alpha": 0.3417365166940647,
|
| 91 |
+
"num_train_epochs": 8,
|
| 92 |
+
"temperature": 2
|
| 93 |
}
|
| 94 |
}
|
run-0/checkpoint-2000/trainer_state.json
CHANGED
|
@@ -10,91 +10,91 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
-
"eval_accuracy": 0.
|
| 14 |
-
"eval_loss": 0.
|
| 15 |
-
"eval_runtime":
|
| 16 |
-
"eval_samples_per_second":
|
| 17 |
-
"eval_steps_per_second":
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
-
"grad_norm": 0.
|
| 23 |
-
"learning_rate": 1.
|
| 24 |
-
"loss": 0.
|
| 25 |
"step": 500
|
| 26 |
},
|
| 27 |
{
|
| 28 |
"epoch": 2.0,
|
| 29 |
-
"eval_accuracy": 0.
|
| 30 |
-
"eval_loss": 0.
|
| 31 |
-
"eval_runtime":
|
| 32 |
-
"eval_samples_per_second":
|
| 33 |
-
"eval_steps_per_second": 10.
|
| 34 |
"step": 636
|
| 35 |
},
|
| 36 |
{
|
| 37 |
"epoch": 3.0,
|
| 38 |
-
"eval_accuracy": 0.
|
| 39 |
-
"eval_loss": 0.
|
| 40 |
-
"eval_runtime": 5.
|
| 41 |
-
"eval_samples_per_second":
|
| 42 |
-
"eval_steps_per_second": 11.
|
| 43 |
"step": 954
|
| 44 |
},
|
| 45 |
{
|
| 46 |
"epoch": 3.1446540880503147,
|
| 47 |
-
"grad_norm": 0.
|
| 48 |
-
"learning_rate": 1.
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 1000
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 4.0,
|
| 54 |
-
"eval_accuracy": 0.
|
| 55 |
-
"eval_loss": 0.
|
| 56 |
-
"eval_runtime": 5.
|
| 57 |
-
"eval_samples_per_second":
|
| 58 |
-
"eval_steps_per_second": 11.
|
| 59 |
"step": 1272
|
| 60 |
},
|
| 61 |
{
|
| 62 |
"epoch": 4.716981132075472,
|
| 63 |
-
"grad_norm": 0.
|
| 64 |
-
"learning_rate":
|
| 65 |
-
"loss": 0.
|
| 66 |
"step": 1500
|
| 67 |
},
|
| 68 |
{
|
| 69 |
"epoch": 5.0,
|
| 70 |
-
"eval_accuracy": 0.
|
| 71 |
-
"eval_loss": 0.
|
| 72 |
-
"eval_runtime":
|
| 73 |
-
"eval_samples_per_second":
|
| 74 |
-
"eval_steps_per_second":
|
| 75 |
"step": 1590
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 6.0,
|
| 79 |
-
"eval_accuracy": 0.
|
| 80 |
-
"eval_loss": 0.
|
| 81 |
-
"eval_runtime": 5.
|
| 82 |
-
"eval_samples_per_second":
|
| 83 |
-
"eval_steps_per_second": 11.
|
| 84 |
"step": 1908
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 6.289308176100629,
|
| 88 |
-
"grad_norm": 0.
|
| 89 |
-
"learning_rate":
|
| 90 |
-
"loss": 0.
|
| 91 |
"step": 2000
|
| 92 |
}
|
| 93 |
],
|
| 94 |
"logging_steps": 500,
|
| 95 |
-
"max_steps":
|
| 96 |
"num_input_tokens_seen": 0,
|
| 97 |
-
"num_train_epochs":
|
| 98 |
"save_steps": 500,
|
| 99 |
"stateful_callbacks": {
|
| 100 |
"TrainerControl": {
|
|
@@ -108,12 +108,12 @@
|
|
| 108 |
"attributes": {}
|
| 109 |
}
|
| 110 |
},
|
| 111 |
-
"total_flos":
|
| 112 |
"train_batch_size": 48,
|
| 113 |
"trial_name": null,
|
| 114 |
"trial_params": {
|
| 115 |
-
"alpha": 0.
|
| 116 |
-
"num_train_epochs":
|
| 117 |
-
"temperature":
|
| 118 |
}
|
| 119 |
}
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.6758064516129032,
|
| 14 |
+
"eval_loss": 0.40983492136001587,
|
| 15 |
+
"eval_runtime": 5.8092,
|
| 16 |
+
"eval_samples_per_second": 533.639,
|
| 17 |
+
"eval_steps_per_second": 11.189,
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.86518794298172,
|
| 23 |
+
"learning_rate": 1.606918238993711e-05,
|
| 24 |
+
"loss": 0.645,
|
| 25 |
"step": 500
|
| 26 |
},
|
| 27 |
{
|
| 28 |
"epoch": 2.0,
|
| 29 |
+
"eval_accuracy": 0.8412903225806452,
|
| 30 |
+
"eval_loss": 0.142293781042099,
|
| 31 |
+
"eval_runtime": 6.1991,
|
| 32 |
+
"eval_samples_per_second": 500.073,
|
| 33 |
+
"eval_steps_per_second": 10.485,
|
| 34 |
"step": 636
|
| 35 |
},
|
| 36 |
{
|
| 37 |
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.8938709677419355,
|
| 39 |
+
"eval_loss": 0.07486626505851746,
|
| 40 |
+
"eval_runtime": 5.5967,
|
| 41 |
+
"eval_samples_per_second": 553.895,
|
| 42 |
+
"eval_steps_per_second": 11.614,
|
| 43 |
"step": 954
|
| 44 |
},
|
| 45 |
{
|
| 46 |
"epoch": 3.1446540880503147,
|
| 47 |
+
"grad_norm": 0.7068584561347961,
|
| 48 |
+
"learning_rate": 1.2138364779874214e-05,
|
| 49 |
+
"loss": 0.1671,
|
| 50 |
"step": 1000
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 4.0,
|
| 54 |
+
"eval_accuracy": 0.9141935483870968,
|
| 55 |
+
"eval_loss": 0.053671687841415405,
|
| 56 |
+
"eval_runtime": 5.7734,
|
| 57 |
+
"eval_samples_per_second": 536.949,
|
| 58 |
+
"eval_steps_per_second": 11.259,
|
| 59 |
"step": 1272
|
| 60 |
},
|
| 61 |
{
|
| 62 |
"epoch": 4.716981132075472,
|
| 63 |
+
"grad_norm": 0.464497447013855,
|
| 64 |
+
"learning_rate": 8.207547169811321e-06,
|
| 65 |
+
"loss": 0.0849,
|
| 66 |
"step": 1500
|
| 67 |
},
|
| 68 |
{
|
| 69 |
"epoch": 5.0,
|
| 70 |
+
"eval_accuracy": 0.9238709677419354,
|
| 71 |
+
"eval_loss": 0.04483605548739433,
|
| 72 |
+
"eval_runtime": 6.0581,
|
| 73 |
+
"eval_samples_per_second": 511.712,
|
| 74 |
+
"eval_steps_per_second": 10.729,
|
| 75 |
"step": 1590
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 6.0,
|
| 79 |
+
"eval_accuracy": 0.9261290322580645,
|
| 80 |
+
"eval_loss": 0.04038877785205841,
|
| 81 |
+
"eval_runtime": 5.6704,
|
| 82 |
+
"eval_samples_per_second": 546.695,
|
| 83 |
+
"eval_steps_per_second": 11.463,
|
| 84 |
"step": 1908
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 6.289308176100629,
|
| 88 |
+
"grad_norm": 0.36644139885902405,
|
| 89 |
+
"learning_rate": 4.276729559748428e-06,
|
| 90 |
+
"loss": 0.0648,
|
| 91 |
"step": 2000
|
| 92 |
}
|
| 93 |
],
|
| 94 |
"logging_steps": 500,
|
| 95 |
+
"max_steps": 2544,
|
| 96 |
"num_input_tokens_seen": 0,
|
| 97 |
+
"num_train_epochs": 8,
|
| 98 |
"save_steps": 500,
|
| 99 |
"stateful_callbacks": {
|
| 100 |
"TrainerControl": {
|
|
|
|
| 108 |
"attributes": {}
|
| 109 |
}
|
| 110 |
},
|
| 111 |
+
"total_flos": 568507463976984.0,
|
| 112 |
"train_batch_size": 48,
|
| 113 |
"trial_name": null,
|
| 114 |
"trial_params": {
|
| 115 |
+
"alpha": 0.3417365166940647,
|
| 116 |
+
"num_train_epochs": 8,
|
| 117 |
+
"temperature": 2
|
| 118 |
}
|
| 119 |
}
|
run-0/checkpoint-500/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 268290900
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1cdf92297df9f81ffbfd1ae3766b8e921521ef797e1630b0e924193535c5bc0
|
| 3 |
size 268290900
|
run-0/checkpoint-500/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 536643898
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa58f4b006c0c28a3daff33cb29801956213107477066afa431acfa26604cfb8
|
| 3 |
size 536643898
|
run-0/checkpoint-500/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90dc4637e972cc69c745eebddd8a7560dca27d2318df3e23f8e145abbf236536
|
| 3 |
size 1064
|
run-0/checkpoint-500/trainer_state.json
CHANGED
|
@@ -10,25 +10,25 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
-
"eval_accuracy": 0.
|
| 14 |
-
"eval_loss": 0.
|
| 15 |
-
"eval_runtime":
|
| 16 |
-
"eval_samples_per_second":
|
| 17 |
-
"eval_steps_per_second":
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
-
"grad_norm": 0.
|
| 23 |
-
"learning_rate": 1.
|
| 24 |
-
"loss": 0.
|
| 25 |
"step": 500
|
| 26 |
}
|
| 27 |
],
|
| 28 |
"logging_steps": 500,
|
| 29 |
-
"max_steps":
|
| 30 |
"num_input_tokens_seen": 0,
|
| 31 |
-
"num_train_epochs":
|
| 32 |
"save_steps": 500,
|
| 33 |
"stateful_callbacks": {
|
| 34 |
"TrainerControl": {
|
|
@@ -42,12 +42,12 @@
|
|
| 42 |
"attributes": {}
|
| 43 |
}
|
| 44 |
},
|
| 45 |
-
"total_flos":
|
| 46 |
"train_batch_size": 48,
|
| 47 |
"trial_name": null,
|
| 48 |
"trial_params": {
|
| 49 |
-
"alpha": 0.
|
| 50 |
-
"num_train_epochs":
|
| 51 |
-
"temperature":
|
| 52 |
}
|
| 53 |
}
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.6758064516129032,
|
| 14 |
+
"eval_loss": 0.40983492136001587,
|
| 15 |
+
"eval_runtime": 5.8092,
|
| 16 |
+
"eval_samples_per_second": 533.639,
|
| 17 |
+
"eval_steps_per_second": 11.189,
|
| 18 |
"step": 318
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.86518794298172,
|
| 23 |
+
"learning_rate": 1.606918238993711e-05,
|
| 24 |
+
"loss": 0.645,
|
| 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": {
|
|
|
|
| 42 |
"attributes": {}
|
| 43 |
}
|
| 44 |
},
|
| 45 |
+
"total_flos": 177588346457172.0,
|
| 46 |
"train_batch_size": 48,
|
| 47 |
"trial_name": null,
|
| 48 |
"trial_params": {
|
| 49 |
+
"alpha": 0.3417365166940647,
|
| 50 |
+
"num_train_epochs": 8,
|
| 51 |
+
"temperature": 2
|
| 52 |
}
|
| 53 |
}
|
run-0/checkpoint-500/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a795e38fc17d23ea7e491b0b97849dc60c777dbce157fcfdc837d87b44f51896
|
| 3 |
size 5176
|
run-7/checkpoint-2500/config.json
ADDED
|
@@ -0,0 +1,331 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert-base-uncased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DistilBertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "restaurant_reviews",
|
| 13 |
+
"1": "nutrition_info",
|
| 14 |
+
"2": "account_blocked",
|
| 15 |
+
"3": "oil_change_how",
|
| 16 |
+
"4": "time",
|
| 17 |
+
"5": "weather",
|
| 18 |
+
"6": "redeem_rewards",
|
| 19 |
+
"7": "interest_rate",
|
| 20 |
+
"8": "gas_type",
|
| 21 |
+
"9": "accept_reservations",
|
| 22 |
+
"10": "smart_home",
|
| 23 |
+
"11": "user_name",
|
| 24 |
+
"12": "report_lost_card",
|
| 25 |
+
"13": "repeat",
|
| 26 |
+
"14": "whisper_mode",
|
| 27 |
+
"15": "what_are_your_hobbies",
|
| 28 |
+
"16": "order",
|
| 29 |
+
"17": "jump_start",
|
| 30 |
+
"18": "schedule_meeting",
|
| 31 |
+
"19": "meeting_schedule",
|
| 32 |
+
"20": "freeze_account",
|
| 33 |
+
"21": "what_song",
|
| 34 |
+
"22": "meaning_of_life",
|
| 35 |
+
"23": "restaurant_reservation",
|
| 36 |
+
"24": "traffic",
|
| 37 |
+
"25": "make_call",
|
| 38 |
+
"26": "text",
|
| 39 |
+
"27": "bill_balance",
|
| 40 |
+
"28": "improve_credit_score",
|
| 41 |
+
"29": "change_language",
|
| 42 |
+
"30": "no",
|
| 43 |
+
"31": "measurement_conversion",
|
| 44 |
+
"32": "timer",
|
| 45 |
+
"33": "flip_coin",
|
| 46 |
+
"34": "do_you_have_pets",
|
| 47 |
+
"35": "balance",
|
| 48 |
+
"36": "tell_joke",
|
| 49 |
+
"37": "last_maintenance",
|
| 50 |
+
"38": "exchange_rate",
|
| 51 |
+
"39": "uber",
|
| 52 |
+
"40": "car_rental",
|
| 53 |
+
"41": "credit_limit",
|
| 54 |
+
"42": "oos",
|
| 55 |
+
"43": "shopping_list",
|
| 56 |
+
"44": "expiration_date",
|
| 57 |
+
"45": "routing",
|
| 58 |
+
"46": "meal_suggestion",
|
| 59 |
+
"47": "tire_change",
|
| 60 |
+
"48": "todo_list",
|
| 61 |
+
"49": "card_declined",
|
| 62 |
+
"50": "rewards_balance",
|
| 63 |
+
"51": "change_accent",
|
| 64 |
+
"52": "vaccines",
|
| 65 |
+
"53": "reminder_update",
|
| 66 |
+
"54": "food_last",
|
| 67 |
+
"55": "change_ai_name",
|
| 68 |
+
"56": "bill_due",
|
| 69 |
+
"57": "who_do_you_work_for",
|
| 70 |
+
"58": "share_location",
|
| 71 |
+
"59": "international_visa",
|
| 72 |
+
"60": "calendar",
|
| 73 |
+
"61": "translate",
|
| 74 |
+
"62": "carry_on",
|
| 75 |
+
"63": "book_flight",
|
| 76 |
+
"64": "insurance_change",
|
| 77 |
+
"65": "todo_list_update",
|
| 78 |
+
"66": "timezone",
|
| 79 |
+
"67": "cancel_reservation",
|
| 80 |
+
"68": "transactions",
|
| 81 |
+
"69": "credit_score",
|
| 82 |
+
"70": "report_fraud",
|
| 83 |
+
"71": "spending_history",
|
| 84 |
+
"72": "directions",
|
| 85 |
+
"73": "spelling",
|
| 86 |
+
"74": "insurance",
|
| 87 |
+
"75": "what_is_your_name",
|
| 88 |
+
"76": "reminder",
|
| 89 |
+
"77": "where_are_you_from",
|
| 90 |
+
"78": "distance",
|
| 91 |
+
"79": "payday",
|
| 92 |
+
"80": "flight_status",
|
| 93 |
+
"81": "find_phone",
|
| 94 |
+
"82": "greeting",
|
| 95 |
+
"83": "alarm",
|
| 96 |
+
"84": "order_status",
|
| 97 |
+
"85": "confirm_reservation",
|
| 98 |
+
"86": "cook_time",
|
| 99 |
+
"87": "damaged_card",
|
| 100 |
+
"88": "reset_settings",
|
| 101 |
+
"89": "pin_change",
|
| 102 |
+
"90": "replacement_card_duration",
|
| 103 |
+
"91": "new_card",
|
| 104 |
+
"92": "roll_dice",
|
| 105 |
+
"93": "income",
|
| 106 |
+
"94": "taxes",
|
| 107 |
+
"95": "date",
|
| 108 |
+
"96": "who_made_you",
|
| 109 |
+
"97": "pto_request",
|
| 110 |
+
"98": "tire_pressure",
|
| 111 |
+
"99": "how_old_are_you",
|
| 112 |
+
"100": "rollover_401k",
|
| 113 |
+
"101": "pto_request_status",
|
| 114 |
+
"102": "how_busy",
|
| 115 |
+
"103": "application_status",
|
| 116 |
+
"104": "recipe",
|
| 117 |
+
"105": "calendar_update",
|
| 118 |
+
"106": "play_music",
|
| 119 |
+
"107": "yes",
|
| 120 |
+
"108": "direct_deposit",
|
| 121 |
+
"109": "credit_limit_change",
|
| 122 |
+
"110": "gas",
|
| 123 |
+
"111": "pay_bill",
|
| 124 |
+
"112": "ingredients_list",
|
| 125 |
+
"113": "lost_luggage",
|
| 126 |
+
"114": "goodbye",
|
| 127 |
+
"115": "what_can_i_ask_you",
|
| 128 |
+
"116": "book_hotel",
|
| 129 |
+
"117": "are_you_a_bot",
|
| 130 |
+
"118": "next_song",
|
| 131 |
+
"119": "change_speed",
|
| 132 |
+
"120": "plug_type",
|
| 133 |
+
"121": "maybe",
|
| 134 |
+
"122": "w2",
|
| 135 |
+
"123": "oil_change_when",
|
| 136 |
+
"124": "thank_you",
|
| 137 |
+
"125": "shopping_list_update",
|
| 138 |
+
"126": "pto_balance",
|
| 139 |
+
"127": "order_checks",
|
| 140 |
+
"128": "travel_alert",
|
| 141 |
+
"129": "fun_fact",
|
| 142 |
+
"130": "sync_device",
|
| 143 |
+
"131": "schedule_maintenance",
|
| 144 |
+
"132": "apr",
|
| 145 |
+
"133": "transfer",
|
| 146 |
+
"134": "ingredient_substitution",
|
| 147 |
+
"135": "calories",
|
| 148 |
+
"136": "current_location",
|
| 149 |
+
"137": "international_fees",
|
| 150 |
+
"138": "calculator",
|
| 151 |
+
"139": "definition",
|
| 152 |
+
"140": "next_holiday",
|
| 153 |
+
"141": "update_playlist",
|
| 154 |
+
"142": "mpg",
|
| 155 |
+
"143": "min_payment",
|
| 156 |
+
"144": "change_user_name",
|
| 157 |
+
"145": "restaurant_suggestion",
|
| 158 |
+
"146": "travel_notification",
|
| 159 |
+
"147": "cancel",
|
| 160 |
+
"148": "pto_used",
|
| 161 |
+
"149": "travel_suggestion",
|
| 162 |
+
"150": "change_volume"
|
| 163 |
+
},
|
| 164 |
+
"initializer_range": 0.02,
|
| 165 |
+
"label2id": {
|
| 166 |
+
"accept_reservations": 9,
|
| 167 |
+
"account_blocked": 2,
|
| 168 |
+
"alarm": 83,
|
| 169 |
+
"application_status": 103,
|
| 170 |
+
"apr": 132,
|
| 171 |
+
"are_you_a_bot": 117,
|
| 172 |
+
"balance": 35,
|
| 173 |
+
"bill_balance": 27,
|
| 174 |
+
"bill_due": 56,
|
| 175 |
+
"book_flight": 63,
|
| 176 |
+
"book_hotel": 116,
|
| 177 |
+
"calculator": 138,
|
| 178 |
+
"calendar": 60,
|
| 179 |
+
"calendar_update": 105,
|
| 180 |
+
"calories": 135,
|
| 181 |
+
"cancel": 147,
|
| 182 |
+
"cancel_reservation": 67,
|
| 183 |
+
"car_rental": 40,
|
| 184 |
+
"card_declined": 49,
|
| 185 |
+
"carry_on": 62,
|
| 186 |
+
"change_accent": 51,
|
| 187 |
+
"change_ai_name": 55,
|
| 188 |
+
"change_language": 29,
|
| 189 |
+
"change_speed": 119,
|
| 190 |
+
"change_user_name": 144,
|
| 191 |
+
"change_volume": 150,
|
| 192 |
+
"confirm_reservation": 85,
|
| 193 |
+
"cook_time": 86,
|
| 194 |
+
"credit_limit": 41,
|
| 195 |
+
"credit_limit_change": 109,
|
| 196 |
+
"credit_score": 69,
|
| 197 |
+
"current_location": 136,
|
| 198 |
+
"damaged_card": 87,
|
| 199 |
+
"date": 95,
|
| 200 |
+
"definition": 139,
|
| 201 |
+
"direct_deposit": 108,
|
| 202 |
+
"directions": 72,
|
| 203 |
+
"distance": 78,
|
| 204 |
+
"do_you_have_pets": 34,
|
| 205 |
+
"exchange_rate": 38,
|
| 206 |
+
"expiration_date": 44,
|
| 207 |
+
"find_phone": 81,
|
| 208 |
+
"flight_status": 80,
|
| 209 |
+
"flip_coin": 33,
|
| 210 |
+
"food_last": 54,
|
| 211 |
+
"freeze_account": 20,
|
| 212 |
+
"fun_fact": 129,
|
| 213 |
+
"gas": 110,
|
| 214 |
+
"gas_type": 8,
|
| 215 |
+
"goodbye": 114,
|
| 216 |
+
"greeting": 82,
|
| 217 |
+
"how_busy": 102,
|
| 218 |
+
"how_old_are_you": 99,
|
| 219 |
+
"improve_credit_score": 28,
|
| 220 |
+
"income": 93,
|
| 221 |
+
"ingredient_substitution": 134,
|
| 222 |
+
"ingredients_list": 112,
|
| 223 |
+
"insurance": 74,
|
| 224 |
+
"insurance_change": 64,
|
| 225 |
+
"interest_rate": 7,
|
| 226 |
+
"international_fees": 137,
|
| 227 |
+
"international_visa": 59,
|
| 228 |
+
"jump_start": 17,
|
| 229 |
+
"last_maintenance": 37,
|
| 230 |
+
"lost_luggage": 113,
|
| 231 |
+
"make_call": 25,
|
| 232 |
+
"maybe": 121,
|
| 233 |
+
"meal_suggestion": 46,
|
| 234 |
+
"meaning_of_life": 22,
|
| 235 |
+
"measurement_conversion": 31,
|
| 236 |
+
"meeting_schedule": 19,
|
| 237 |
+
"min_payment": 143,
|
| 238 |
+
"mpg": 142,
|
| 239 |
+
"new_card": 91,
|
| 240 |
+
"next_holiday": 140,
|
| 241 |
+
"next_song": 118,
|
| 242 |
+
"no": 30,
|
| 243 |
+
"nutrition_info": 1,
|
| 244 |
+
"oil_change_how": 3,
|
| 245 |
+
"oil_change_when": 123,
|
| 246 |
+
"oos": 42,
|
| 247 |
+
"order": 16,
|
| 248 |
+
"order_checks": 127,
|
| 249 |
+
"order_status": 84,
|
| 250 |
+
"pay_bill": 111,
|
| 251 |
+
"payday": 79,
|
| 252 |
+
"pin_change": 89,
|
| 253 |
+
"play_music": 106,
|
| 254 |
+
"plug_type": 120,
|
| 255 |
+
"pto_balance": 126,
|
| 256 |
+
"pto_request": 97,
|
| 257 |
+
"pto_request_status": 101,
|
| 258 |
+
"pto_used": 148,
|
| 259 |
+
"recipe": 104,
|
| 260 |
+
"redeem_rewards": 6,
|
| 261 |
+
"reminder": 76,
|
| 262 |
+
"reminder_update": 53,
|
| 263 |
+
"repeat": 13,
|
| 264 |
+
"replacement_card_duration": 90,
|
| 265 |
+
"report_fraud": 70,
|
| 266 |
+
"report_lost_card": 12,
|
| 267 |
+
"reset_settings": 88,
|
| 268 |
+
"restaurant_reservation": 23,
|
| 269 |
+
"restaurant_reviews": 0,
|
| 270 |
+
"restaurant_suggestion": 145,
|
| 271 |
+
"rewards_balance": 50,
|
| 272 |
+
"roll_dice": 92,
|
| 273 |
+
"rollover_401k": 100,
|
| 274 |
+
"routing": 45,
|
| 275 |
+
"schedule_maintenance": 131,
|
| 276 |
+
"schedule_meeting": 18,
|
| 277 |
+
"share_location": 58,
|
| 278 |
+
"shopping_list": 43,
|
| 279 |
+
"shopping_list_update": 125,
|
| 280 |
+
"smart_home": 10,
|
| 281 |
+
"spelling": 73,
|
| 282 |
+
"spending_history": 71,
|
| 283 |
+
"sync_device": 130,
|
| 284 |
+
"taxes": 94,
|
| 285 |
+
"tell_joke": 36,
|
| 286 |
+
"text": 26,
|
| 287 |
+
"thank_you": 124,
|
| 288 |
+
"time": 4,
|
| 289 |
+
"timer": 32,
|
| 290 |
+
"timezone": 66,
|
| 291 |
+
"tire_change": 47,
|
| 292 |
+
"tire_pressure": 98,
|
| 293 |
+
"todo_list": 48,
|
| 294 |
+
"todo_list_update": 65,
|
| 295 |
+
"traffic": 24,
|
| 296 |
+
"transactions": 68,
|
| 297 |
+
"transfer": 133,
|
| 298 |
+
"translate": 61,
|
| 299 |
+
"travel_alert": 128,
|
| 300 |
+
"travel_notification": 146,
|
| 301 |
+
"travel_suggestion": 149,
|
| 302 |
+
"uber": 39,
|
| 303 |
+
"update_playlist": 141,
|
| 304 |
+
"user_name": 11,
|
| 305 |
+
"vaccines": 52,
|
| 306 |
+
"w2": 122,
|
| 307 |
+
"weather": 5,
|
| 308 |
+
"what_are_your_hobbies": 15,
|
| 309 |
+
"what_can_i_ask_you": 115,
|
| 310 |
+
"what_is_your_name": 75,
|
| 311 |
+
"what_song": 21,
|
| 312 |
+
"where_are_you_from": 77,
|
| 313 |
+
"whisper_mode": 14,
|
| 314 |
+
"who_do_you_work_for": 57,
|
| 315 |
+
"who_made_you": 96,
|
| 316 |
+
"yes": 107
|
| 317 |
+
},
|
| 318 |
+
"max_position_embeddings": 512,
|
| 319 |
+
"model_type": "distilbert",
|
| 320 |
+
"n_heads": 12,
|
| 321 |
+
"n_layers": 6,
|
| 322 |
+
"pad_token_id": 0,
|
| 323 |
+
"problem_type": "single_label_classification",
|
| 324 |
+
"qa_dropout": 0.1,
|
| 325 |
+
"seq_classif_dropout": 0.2,
|
| 326 |
+
"sinusoidal_pos_embds": false,
|
| 327 |
+
"tie_weights_": true,
|
| 328 |
+
"torch_dtype": "float32",
|
| 329 |
+
"transformers_version": "4.42.4",
|
| 330 |
+
"vocab_size": 30522
|
| 331 |
+
}
|
run-7/checkpoint-2500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b94f5342d01b1bd65a8c12888edd707e0e2031a89e7cd12ec3ea0b85a904cbe5
|
| 3 |
+
size 268290900
|
run-7/checkpoint-2500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e49385fc343717e4317e1faede64c96d7c46392952318f1e977eb6e6527ada8d
|
| 3 |
+
size 536643898
|
run-7/checkpoint-2500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47dadd34efe7a8b1be3b9e277f2872fb36f6eb0ed07cc82c6991ea5264f9ad54
|
| 3 |
+
size 14244
|
run-7/checkpoint-2500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7c943a77651a184c6653dd4706c14c9a27d52136e81942a2b361c74acb9f665
|
| 3 |
+
size 1064
|
run-7/checkpoint-2500/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
run-7/checkpoint-2500/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
run-7/checkpoint-2500/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "DistilBertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
run-7/checkpoint-2500/trainer_state.json
ADDED
|
@@ -0,0 +1,135 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 7.861635220125786,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2500,
|
| 7 |
+
"is_hyper_param_search": true,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.6,
|
| 14 |
+
"eval_loss": 0.21216703951358795,
|
| 15 |
+
"eval_runtime": 5.8214,
|
| 16 |
+
"eval_samples_per_second": 532.514,
|
| 17 |
+
"eval_steps_per_second": 11.166,
|
| 18 |
+
"step": 318
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.5472351908683777,
|
| 23 |
+
"learning_rate": 1.650593990216632e-05,
|
| 24 |
+
"loss": 0.3411,
|
| 25 |
+
"step": 500
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 2.0,
|
| 29 |
+
"eval_accuracy": 0.8267741935483871,
|
| 30 |
+
"eval_loss": 0.10011853277683258,
|
| 31 |
+
"eval_runtime": 6.0665,
|
| 32 |
+
"eval_samples_per_second": 511.002,
|
| 33 |
+
"eval_steps_per_second": 10.715,
|
| 34 |
+
"step": 636
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.8829032258064516,
|
| 39 |
+
"eval_loss": 0.06541892141103745,
|
| 40 |
+
"eval_runtime": 5.7333,
|
| 41 |
+
"eval_samples_per_second": 540.702,
|
| 42 |
+
"eval_steps_per_second": 11.337,
|
| 43 |
+
"step": 954
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.1446540880503147,
|
| 47 |
+
"grad_norm": 0.4667431116104126,
|
| 48 |
+
"learning_rate": 1.3011879804332637e-05,
|
| 49 |
+
"loss": 0.1165,
|
| 50 |
+
"step": 1000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 4.0,
|
| 54 |
+
"eval_accuracy": 0.9003225806451612,
|
| 55 |
+
"eval_loss": 0.04985565319657326,
|
| 56 |
+
"eval_runtime": 5.7103,
|
| 57 |
+
"eval_samples_per_second": 542.881,
|
| 58 |
+
"eval_steps_per_second": 11.383,
|
| 59 |
+
"step": 1272
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 4.716981132075472,
|
| 63 |
+
"grad_norm": 0.31565672159194946,
|
| 64 |
+
"learning_rate": 9.517819706498952e-06,
|
| 65 |
+
"loss": 0.073,
|
| 66 |
+
"step": 1500
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 5.0,
|
| 70 |
+
"eval_accuracy": 0.9106451612903226,
|
| 71 |
+
"eval_loss": 0.040586207062006,
|
| 72 |
+
"eval_runtime": 5.7897,
|
| 73 |
+
"eval_samples_per_second": 535.429,
|
| 74 |
+
"eval_steps_per_second": 11.227,
|
| 75 |
+
"step": 1590
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 6.0,
|
| 79 |
+
"eval_accuracy": 0.9129032258064517,
|
| 80 |
+
"eval_loss": 0.035505082458257675,
|
| 81 |
+
"eval_runtime": 5.6656,
|
| 82 |
+
"eval_samples_per_second": 547.161,
|
| 83 |
+
"eval_steps_per_second": 11.473,
|
| 84 |
+
"step": 1908
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 6.289308176100629,
|
| 88 |
+
"grad_norm": 0.2728483974933624,
|
| 89 |
+
"learning_rate": 6.02375960866527e-06,
|
| 90 |
+
"loss": 0.0569,
|
| 91 |
+
"step": 2000
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 7.0,
|
| 95 |
+
"eval_accuracy": 0.9219354838709677,
|
| 96 |
+
"eval_loss": 0.03278331458568573,
|
| 97 |
+
"eval_runtime": 5.9424,
|
| 98 |
+
"eval_samples_per_second": 521.675,
|
| 99 |
+
"eval_steps_per_second": 10.938,
|
| 100 |
+
"step": 2226
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 7.861635220125786,
|
| 104 |
+
"grad_norm": 0.2838101089000702,
|
| 105 |
+
"learning_rate": 2.5296995108315863e-06,
|
| 106 |
+
"loss": 0.0498,
|
| 107 |
+
"step": 2500
|
| 108 |
+
}
|
| 109 |
+
],
|
| 110 |
+
"logging_steps": 500,
|
| 111 |
+
"max_steps": 2862,
|
| 112 |
+
"num_input_tokens_seen": 0,
|
| 113 |
+
"num_train_epochs": 9,
|
| 114 |
+
"save_steps": 500,
|
| 115 |
+
"stateful_callbacks": {
|
| 116 |
+
"TrainerControl": {
|
| 117 |
+
"args": {
|
| 118 |
+
"should_epoch_stop": false,
|
| 119 |
+
"should_evaluate": false,
|
| 120 |
+
"should_log": false,
|
| 121 |
+
"should_save": true,
|
| 122 |
+
"should_training_stop": false
|
| 123 |
+
},
|
| 124 |
+
"attributes": {}
|
| 125 |
+
}
|
| 126 |
+
},
|
| 127 |
+
"total_flos": 734160224101416.0,
|
| 128 |
+
"train_batch_size": 48,
|
| 129 |
+
"trial_name": null,
|
| 130 |
+
"trial_params": {
|
| 131 |
+
"alpha": 0.46921361817552565,
|
| 132 |
+
"num_train_epochs": 9,
|
| 133 |
+
"temperature": 7
|
| 134 |
+
}
|
| 135 |
+
}
|
run-7/checkpoint-2500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f906ae2d10e00292b8865bd9d1a4e0ea88513f339077e0689097ea85b524df3d
|
| 3 |
+
size 5176
|
run-7/checkpoint-2500/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
run-7/checkpoint-2862/config.json
ADDED
|
@@ -0,0 +1,331 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert-base-uncased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DistilBertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "restaurant_reviews",
|
| 13 |
+
"1": "nutrition_info",
|
| 14 |
+
"2": "account_blocked",
|
| 15 |
+
"3": "oil_change_how",
|
| 16 |
+
"4": "time",
|
| 17 |
+
"5": "weather",
|
| 18 |
+
"6": "redeem_rewards",
|
| 19 |
+
"7": "interest_rate",
|
| 20 |
+
"8": "gas_type",
|
| 21 |
+
"9": "accept_reservations",
|
| 22 |
+
"10": "smart_home",
|
| 23 |
+
"11": "user_name",
|
| 24 |
+
"12": "report_lost_card",
|
| 25 |
+
"13": "repeat",
|
| 26 |
+
"14": "whisper_mode",
|
| 27 |
+
"15": "what_are_your_hobbies",
|
| 28 |
+
"16": "order",
|
| 29 |
+
"17": "jump_start",
|
| 30 |
+
"18": "schedule_meeting",
|
| 31 |
+
"19": "meeting_schedule",
|
| 32 |
+
"20": "freeze_account",
|
| 33 |
+
"21": "what_song",
|
| 34 |
+
"22": "meaning_of_life",
|
| 35 |
+
"23": "restaurant_reservation",
|
| 36 |
+
"24": "traffic",
|
| 37 |
+
"25": "make_call",
|
| 38 |
+
"26": "text",
|
| 39 |
+
"27": "bill_balance",
|
| 40 |
+
"28": "improve_credit_score",
|
| 41 |
+
"29": "change_language",
|
| 42 |
+
"30": "no",
|
| 43 |
+
"31": "measurement_conversion",
|
| 44 |
+
"32": "timer",
|
| 45 |
+
"33": "flip_coin",
|
| 46 |
+
"34": "do_you_have_pets",
|
| 47 |
+
"35": "balance",
|
| 48 |
+
"36": "tell_joke",
|
| 49 |
+
"37": "last_maintenance",
|
| 50 |
+
"38": "exchange_rate",
|
| 51 |
+
"39": "uber",
|
| 52 |
+
"40": "car_rental",
|
| 53 |
+
"41": "credit_limit",
|
| 54 |
+
"42": "oos",
|
| 55 |
+
"43": "shopping_list",
|
| 56 |
+
"44": "expiration_date",
|
| 57 |
+
"45": "routing",
|
| 58 |
+
"46": "meal_suggestion",
|
| 59 |
+
"47": "tire_change",
|
| 60 |
+
"48": "todo_list",
|
| 61 |
+
"49": "card_declined",
|
| 62 |
+
"50": "rewards_balance",
|
| 63 |
+
"51": "change_accent",
|
| 64 |
+
"52": "vaccines",
|
| 65 |
+
"53": "reminder_update",
|
| 66 |
+
"54": "food_last",
|
| 67 |
+
"55": "change_ai_name",
|
| 68 |
+
"56": "bill_due",
|
| 69 |
+
"57": "who_do_you_work_for",
|
| 70 |
+
"58": "share_location",
|
| 71 |
+
"59": "international_visa",
|
| 72 |
+
"60": "calendar",
|
| 73 |
+
"61": "translate",
|
| 74 |
+
"62": "carry_on",
|
| 75 |
+
"63": "book_flight",
|
| 76 |
+
"64": "insurance_change",
|
| 77 |
+
"65": "todo_list_update",
|
| 78 |
+
"66": "timezone",
|
| 79 |
+
"67": "cancel_reservation",
|
| 80 |
+
"68": "transactions",
|
| 81 |
+
"69": "credit_score",
|
| 82 |
+
"70": "report_fraud",
|
| 83 |
+
"71": "spending_history",
|
| 84 |
+
"72": "directions",
|
| 85 |
+
"73": "spelling",
|
| 86 |
+
"74": "insurance",
|
| 87 |
+
"75": "what_is_your_name",
|
| 88 |
+
"76": "reminder",
|
| 89 |
+
"77": "where_are_you_from",
|
| 90 |
+
"78": "distance",
|
| 91 |
+
"79": "payday",
|
| 92 |
+
"80": "flight_status",
|
| 93 |
+
"81": "find_phone",
|
| 94 |
+
"82": "greeting",
|
| 95 |
+
"83": "alarm",
|
| 96 |
+
"84": "order_status",
|
| 97 |
+
"85": "confirm_reservation",
|
| 98 |
+
"86": "cook_time",
|
| 99 |
+
"87": "damaged_card",
|
| 100 |
+
"88": "reset_settings",
|
| 101 |
+
"89": "pin_change",
|
| 102 |
+
"90": "replacement_card_duration",
|
| 103 |
+
"91": "new_card",
|
| 104 |
+
"92": "roll_dice",
|
| 105 |
+
"93": "income",
|
| 106 |
+
"94": "taxes",
|
| 107 |
+
"95": "date",
|
| 108 |
+
"96": "who_made_you",
|
| 109 |
+
"97": "pto_request",
|
| 110 |
+
"98": "tire_pressure",
|
| 111 |
+
"99": "how_old_are_you",
|
| 112 |
+
"100": "rollover_401k",
|
| 113 |
+
"101": "pto_request_status",
|
| 114 |
+
"102": "how_busy",
|
| 115 |
+
"103": "application_status",
|
| 116 |
+
"104": "recipe",
|
| 117 |
+
"105": "calendar_update",
|
| 118 |
+
"106": "play_music",
|
| 119 |
+
"107": "yes",
|
| 120 |
+
"108": "direct_deposit",
|
| 121 |
+
"109": "credit_limit_change",
|
| 122 |
+
"110": "gas",
|
| 123 |
+
"111": "pay_bill",
|
| 124 |
+
"112": "ingredients_list",
|
| 125 |
+
"113": "lost_luggage",
|
| 126 |
+
"114": "goodbye",
|
| 127 |
+
"115": "what_can_i_ask_you",
|
| 128 |
+
"116": "book_hotel",
|
| 129 |
+
"117": "are_you_a_bot",
|
| 130 |
+
"118": "next_song",
|
| 131 |
+
"119": "change_speed",
|
| 132 |
+
"120": "plug_type",
|
| 133 |
+
"121": "maybe",
|
| 134 |
+
"122": "w2",
|
| 135 |
+
"123": "oil_change_when",
|
| 136 |
+
"124": "thank_you",
|
| 137 |
+
"125": "shopping_list_update",
|
| 138 |
+
"126": "pto_balance",
|
| 139 |
+
"127": "order_checks",
|
| 140 |
+
"128": "travel_alert",
|
| 141 |
+
"129": "fun_fact",
|
| 142 |
+
"130": "sync_device",
|
| 143 |
+
"131": "schedule_maintenance",
|
| 144 |
+
"132": "apr",
|
| 145 |
+
"133": "transfer",
|
| 146 |
+
"134": "ingredient_substitution",
|
| 147 |
+
"135": "calories",
|
| 148 |
+
"136": "current_location",
|
| 149 |
+
"137": "international_fees",
|
| 150 |
+
"138": "calculator",
|
| 151 |
+
"139": "definition",
|
| 152 |
+
"140": "next_holiday",
|
| 153 |
+
"141": "update_playlist",
|
| 154 |
+
"142": "mpg",
|
| 155 |
+
"143": "min_payment",
|
| 156 |
+
"144": "change_user_name",
|
| 157 |
+
"145": "restaurant_suggestion",
|
| 158 |
+
"146": "travel_notification",
|
| 159 |
+
"147": "cancel",
|
| 160 |
+
"148": "pto_used",
|
| 161 |
+
"149": "travel_suggestion",
|
| 162 |
+
"150": "change_volume"
|
| 163 |
+
},
|
| 164 |
+
"initializer_range": 0.02,
|
| 165 |
+
"label2id": {
|
| 166 |
+
"accept_reservations": 9,
|
| 167 |
+
"account_blocked": 2,
|
| 168 |
+
"alarm": 83,
|
| 169 |
+
"application_status": 103,
|
| 170 |
+
"apr": 132,
|
| 171 |
+
"are_you_a_bot": 117,
|
| 172 |
+
"balance": 35,
|
| 173 |
+
"bill_balance": 27,
|
| 174 |
+
"bill_due": 56,
|
| 175 |
+
"book_flight": 63,
|
| 176 |
+
"book_hotel": 116,
|
| 177 |
+
"calculator": 138,
|
| 178 |
+
"calendar": 60,
|
| 179 |
+
"calendar_update": 105,
|
| 180 |
+
"calories": 135,
|
| 181 |
+
"cancel": 147,
|
| 182 |
+
"cancel_reservation": 67,
|
| 183 |
+
"car_rental": 40,
|
| 184 |
+
"card_declined": 49,
|
| 185 |
+
"carry_on": 62,
|
| 186 |
+
"change_accent": 51,
|
| 187 |
+
"change_ai_name": 55,
|
| 188 |
+
"change_language": 29,
|
| 189 |
+
"change_speed": 119,
|
| 190 |
+
"change_user_name": 144,
|
| 191 |
+
"change_volume": 150,
|
| 192 |
+
"confirm_reservation": 85,
|
| 193 |
+
"cook_time": 86,
|
| 194 |
+
"credit_limit": 41,
|
| 195 |
+
"credit_limit_change": 109,
|
| 196 |
+
"credit_score": 69,
|
| 197 |
+
"current_location": 136,
|
| 198 |
+
"damaged_card": 87,
|
| 199 |
+
"date": 95,
|
| 200 |
+
"definition": 139,
|
| 201 |
+
"direct_deposit": 108,
|
| 202 |
+
"directions": 72,
|
| 203 |
+
"distance": 78,
|
| 204 |
+
"do_you_have_pets": 34,
|
| 205 |
+
"exchange_rate": 38,
|
| 206 |
+
"expiration_date": 44,
|
| 207 |
+
"find_phone": 81,
|
| 208 |
+
"flight_status": 80,
|
| 209 |
+
"flip_coin": 33,
|
| 210 |
+
"food_last": 54,
|
| 211 |
+
"freeze_account": 20,
|
| 212 |
+
"fun_fact": 129,
|
| 213 |
+
"gas": 110,
|
| 214 |
+
"gas_type": 8,
|
| 215 |
+
"goodbye": 114,
|
| 216 |
+
"greeting": 82,
|
| 217 |
+
"how_busy": 102,
|
| 218 |
+
"how_old_are_you": 99,
|
| 219 |
+
"improve_credit_score": 28,
|
| 220 |
+
"income": 93,
|
| 221 |
+
"ingredient_substitution": 134,
|
| 222 |
+
"ingredients_list": 112,
|
| 223 |
+
"insurance": 74,
|
| 224 |
+
"insurance_change": 64,
|
| 225 |
+
"interest_rate": 7,
|
| 226 |
+
"international_fees": 137,
|
| 227 |
+
"international_visa": 59,
|
| 228 |
+
"jump_start": 17,
|
| 229 |
+
"last_maintenance": 37,
|
| 230 |
+
"lost_luggage": 113,
|
| 231 |
+
"make_call": 25,
|
| 232 |
+
"maybe": 121,
|
| 233 |
+
"meal_suggestion": 46,
|
| 234 |
+
"meaning_of_life": 22,
|
| 235 |
+
"measurement_conversion": 31,
|
| 236 |
+
"meeting_schedule": 19,
|
| 237 |
+
"min_payment": 143,
|
| 238 |
+
"mpg": 142,
|
| 239 |
+
"new_card": 91,
|
| 240 |
+
"next_holiday": 140,
|
| 241 |
+
"next_song": 118,
|
| 242 |
+
"no": 30,
|
| 243 |
+
"nutrition_info": 1,
|
| 244 |
+
"oil_change_how": 3,
|
| 245 |
+
"oil_change_when": 123,
|
| 246 |
+
"oos": 42,
|
| 247 |
+
"order": 16,
|
| 248 |
+
"order_checks": 127,
|
| 249 |
+
"order_status": 84,
|
| 250 |
+
"pay_bill": 111,
|
| 251 |
+
"payday": 79,
|
| 252 |
+
"pin_change": 89,
|
| 253 |
+
"play_music": 106,
|
| 254 |
+
"plug_type": 120,
|
| 255 |
+
"pto_balance": 126,
|
| 256 |
+
"pto_request": 97,
|
| 257 |
+
"pto_request_status": 101,
|
| 258 |
+
"pto_used": 148,
|
| 259 |
+
"recipe": 104,
|
| 260 |
+
"redeem_rewards": 6,
|
| 261 |
+
"reminder": 76,
|
| 262 |
+
"reminder_update": 53,
|
| 263 |
+
"repeat": 13,
|
| 264 |
+
"replacement_card_duration": 90,
|
| 265 |
+
"report_fraud": 70,
|
| 266 |
+
"report_lost_card": 12,
|
| 267 |
+
"reset_settings": 88,
|
| 268 |
+
"restaurant_reservation": 23,
|
| 269 |
+
"restaurant_reviews": 0,
|
| 270 |
+
"restaurant_suggestion": 145,
|
| 271 |
+
"rewards_balance": 50,
|
| 272 |
+
"roll_dice": 92,
|
| 273 |
+
"rollover_401k": 100,
|
| 274 |
+
"routing": 45,
|
| 275 |
+
"schedule_maintenance": 131,
|
| 276 |
+
"schedule_meeting": 18,
|
| 277 |
+
"share_location": 58,
|
| 278 |
+
"shopping_list": 43,
|
| 279 |
+
"shopping_list_update": 125,
|
| 280 |
+
"smart_home": 10,
|
| 281 |
+
"spelling": 73,
|
| 282 |
+
"spending_history": 71,
|
| 283 |
+
"sync_device": 130,
|
| 284 |
+
"taxes": 94,
|
| 285 |
+
"tell_joke": 36,
|
| 286 |
+
"text": 26,
|
| 287 |
+
"thank_you": 124,
|
| 288 |
+
"time": 4,
|
| 289 |
+
"timer": 32,
|
| 290 |
+
"timezone": 66,
|
| 291 |
+
"tire_change": 47,
|
| 292 |
+
"tire_pressure": 98,
|
| 293 |
+
"todo_list": 48,
|
| 294 |
+
"todo_list_update": 65,
|
| 295 |
+
"traffic": 24,
|
| 296 |
+
"transactions": 68,
|
| 297 |
+
"transfer": 133,
|
| 298 |
+
"translate": 61,
|
| 299 |
+
"travel_alert": 128,
|
| 300 |
+
"travel_notification": 146,
|
| 301 |
+
"travel_suggestion": 149,
|
| 302 |
+
"uber": 39,
|
| 303 |
+
"update_playlist": 141,
|
| 304 |
+
"user_name": 11,
|
| 305 |
+
"vaccines": 52,
|
| 306 |
+
"w2": 122,
|
| 307 |
+
"weather": 5,
|
| 308 |
+
"what_are_your_hobbies": 15,
|
| 309 |
+
"what_can_i_ask_you": 115,
|
| 310 |
+
"what_is_your_name": 75,
|
| 311 |
+
"what_song": 21,
|
| 312 |
+
"where_are_you_from": 77,
|
| 313 |
+
"whisper_mode": 14,
|
| 314 |
+
"who_do_you_work_for": 57,
|
| 315 |
+
"who_made_you": 96,
|
| 316 |
+
"yes": 107
|
| 317 |
+
},
|
| 318 |
+
"max_position_embeddings": 512,
|
| 319 |
+
"model_type": "distilbert",
|
| 320 |
+
"n_heads": 12,
|
| 321 |
+
"n_layers": 6,
|
| 322 |
+
"pad_token_id": 0,
|
| 323 |
+
"problem_type": "single_label_classification",
|
| 324 |
+
"qa_dropout": 0.1,
|
| 325 |
+
"seq_classif_dropout": 0.2,
|
| 326 |
+
"sinusoidal_pos_embds": false,
|
| 327 |
+
"tie_weights_": true,
|
| 328 |
+
"torch_dtype": "float32",
|
| 329 |
+
"transformers_version": "4.42.4",
|
| 330 |
+
"vocab_size": 30522
|
| 331 |
+
}
|
run-7/checkpoint-2862/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25eafb3fe24697973781464a34cc29dc219410843797928327ffd3dcc9539de3
|
| 3 |
+
size 268290900
|
run-7/checkpoint-2862/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c14756d42cc9138f179c383cf379b80870b3a5243e75addd9469864dbfd2bd0
|
| 3 |
+
size 536643898
|
run-7/checkpoint-2862/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb284f354e17b6a985f5b0a1ec258af1589c2ddbdcd17b3670ab73d9e89918e1
|
| 3 |
+
size 14244
|
run-7/checkpoint-2862/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2a738824491ff8b261b6db2a842e15bf80355d04a07892a16df2c4715b2ee556
|
| 3 |
+
size 1064
|
run-7/checkpoint-2862/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
run-7/checkpoint-2862/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
run-7/checkpoint-2862/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "DistilBertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
run-7/checkpoint-2862/trainer_state.json
ADDED
|
@@ -0,0 +1,144 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2862,
|
| 7 |
+
"is_hyper_param_search": true,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.6,
|
| 14 |
+
"eval_loss": 0.21216703951358795,
|
| 15 |
+
"eval_runtime": 5.8214,
|
| 16 |
+
"eval_samples_per_second": 532.514,
|
| 17 |
+
"eval_steps_per_second": 11.166,
|
| 18 |
+
"step": 318
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.5472351908683777,
|
| 23 |
+
"learning_rate": 1.650593990216632e-05,
|
| 24 |
+
"loss": 0.3411,
|
| 25 |
+
"step": 500
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 2.0,
|
| 29 |
+
"eval_accuracy": 0.8267741935483871,
|
| 30 |
+
"eval_loss": 0.10011853277683258,
|
| 31 |
+
"eval_runtime": 6.0665,
|
| 32 |
+
"eval_samples_per_second": 511.002,
|
| 33 |
+
"eval_steps_per_second": 10.715,
|
| 34 |
+
"step": 636
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.8829032258064516,
|
| 39 |
+
"eval_loss": 0.06541892141103745,
|
| 40 |
+
"eval_runtime": 5.7333,
|
| 41 |
+
"eval_samples_per_second": 540.702,
|
| 42 |
+
"eval_steps_per_second": 11.337,
|
| 43 |
+
"step": 954
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.1446540880503147,
|
| 47 |
+
"grad_norm": 0.4667431116104126,
|
| 48 |
+
"learning_rate": 1.3011879804332637e-05,
|
| 49 |
+
"loss": 0.1165,
|
| 50 |
+
"step": 1000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 4.0,
|
| 54 |
+
"eval_accuracy": 0.9003225806451612,
|
| 55 |
+
"eval_loss": 0.04985565319657326,
|
| 56 |
+
"eval_runtime": 5.7103,
|
| 57 |
+
"eval_samples_per_second": 542.881,
|
| 58 |
+
"eval_steps_per_second": 11.383,
|
| 59 |
+
"step": 1272
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 4.716981132075472,
|
| 63 |
+
"grad_norm": 0.31565672159194946,
|
| 64 |
+
"learning_rate": 9.517819706498952e-06,
|
| 65 |
+
"loss": 0.073,
|
| 66 |
+
"step": 1500
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 5.0,
|
| 70 |
+
"eval_accuracy": 0.9106451612903226,
|
| 71 |
+
"eval_loss": 0.040586207062006,
|
| 72 |
+
"eval_runtime": 5.7897,
|
| 73 |
+
"eval_samples_per_second": 535.429,
|
| 74 |
+
"eval_steps_per_second": 11.227,
|
| 75 |
+
"step": 1590
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 6.0,
|
| 79 |
+
"eval_accuracy": 0.9129032258064517,
|
| 80 |
+
"eval_loss": 0.035505082458257675,
|
| 81 |
+
"eval_runtime": 5.6656,
|
| 82 |
+
"eval_samples_per_second": 547.161,
|
| 83 |
+
"eval_steps_per_second": 11.473,
|
| 84 |
+
"step": 1908
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 6.289308176100629,
|
| 88 |
+
"grad_norm": 0.2728483974933624,
|
| 89 |
+
"learning_rate": 6.02375960866527e-06,
|
| 90 |
+
"loss": 0.0569,
|
| 91 |
+
"step": 2000
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 7.0,
|
| 95 |
+
"eval_accuracy": 0.9219354838709677,
|
| 96 |
+
"eval_loss": 0.03278331458568573,
|
| 97 |
+
"eval_runtime": 5.9424,
|
| 98 |
+
"eval_samples_per_second": 521.675,
|
| 99 |
+
"eval_steps_per_second": 10.938,
|
| 100 |
+
"step": 2226
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 7.861635220125786,
|
| 104 |
+
"grad_norm": 0.2838101089000702,
|
| 105 |
+
"learning_rate": 2.5296995108315863e-06,
|
| 106 |
+
"loss": 0.0498,
|
| 107 |
+
"step": 2500
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 8.0,
|
| 111 |
+
"eval_accuracy": 0.9251612903225807,
|
| 112 |
+
"eval_loss": 0.03092290833592415,
|
| 113 |
+
"eval_runtime": 5.7304,
|
| 114 |
+
"eval_samples_per_second": 540.977,
|
| 115 |
+
"eval_steps_per_second": 11.343,
|
| 116 |
+
"step": 2544
|
| 117 |
+
}
|
| 118 |
+
],
|
| 119 |
+
"logging_steps": 500,
|
| 120 |
+
"max_steps": 2862,
|
| 121 |
+
"num_input_tokens_seen": 0,
|
| 122 |
+
"num_train_epochs": 9,
|
| 123 |
+
"save_steps": 500,
|
| 124 |
+
"stateful_callbacks": {
|
| 125 |
+
"TrainerControl": {
|
| 126 |
+
"args": {
|
| 127 |
+
"should_epoch_stop": false,
|
| 128 |
+
"should_evaluate": false,
|
| 129 |
+
"should_log": false,
|
| 130 |
+
"should_save": true,
|
| 131 |
+
"should_training_stop": true
|
| 132 |
+
},
|
| 133 |
+
"attributes": {}
|
| 134 |
+
}
|
| 135 |
+
},
|
| 136 |
+
"total_flos": 734160224101416.0,
|
| 137 |
+
"train_batch_size": 48,
|
| 138 |
+
"trial_name": null,
|
| 139 |
+
"trial_params": {
|
| 140 |
+
"alpha": 0.46921361817552565,
|
| 141 |
+
"num_train_epochs": 9,
|
| 142 |
+
"temperature": 7
|
| 143 |
+
}
|
| 144 |
+
}
|
run-7/checkpoint-2862/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f906ae2d10e00292b8865bd9d1a4e0ea88513f339077e0689097ea85b524df3d
|
| 3 |
+
size 5176
|
run-7/checkpoint-2862/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
run-8/checkpoint-1000/config.json
ADDED
|
@@ -0,0 +1,331 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert-base-uncased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DistilBertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "restaurant_reviews",
|
| 13 |
+
"1": "nutrition_info",
|
| 14 |
+
"2": "account_blocked",
|
| 15 |
+
"3": "oil_change_how",
|
| 16 |
+
"4": "time",
|
| 17 |
+
"5": "weather",
|
| 18 |
+
"6": "redeem_rewards",
|
| 19 |
+
"7": "interest_rate",
|
| 20 |
+
"8": "gas_type",
|
| 21 |
+
"9": "accept_reservations",
|
| 22 |
+
"10": "smart_home",
|
| 23 |
+
"11": "user_name",
|
| 24 |
+
"12": "report_lost_card",
|
| 25 |
+
"13": "repeat",
|
| 26 |
+
"14": "whisper_mode",
|
| 27 |
+
"15": "what_are_your_hobbies",
|
| 28 |
+
"16": "order",
|
| 29 |
+
"17": "jump_start",
|
| 30 |
+
"18": "schedule_meeting",
|
| 31 |
+
"19": "meeting_schedule",
|
| 32 |
+
"20": "freeze_account",
|
| 33 |
+
"21": "what_song",
|
| 34 |
+
"22": "meaning_of_life",
|
| 35 |
+
"23": "restaurant_reservation",
|
| 36 |
+
"24": "traffic",
|
| 37 |
+
"25": "make_call",
|
| 38 |
+
"26": "text",
|
| 39 |
+
"27": "bill_balance",
|
| 40 |
+
"28": "improve_credit_score",
|
| 41 |
+
"29": "change_language",
|
| 42 |
+
"30": "no",
|
| 43 |
+
"31": "measurement_conversion",
|
| 44 |
+
"32": "timer",
|
| 45 |
+
"33": "flip_coin",
|
| 46 |
+
"34": "do_you_have_pets",
|
| 47 |
+
"35": "balance",
|
| 48 |
+
"36": "tell_joke",
|
| 49 |
+
"37": "last_maintenance",
|
| 50 |
+
"38": "exchange_rate",
|
| 51 |
+
"39": "uber",
|
| 52 |
+
"40": "car_rental",
|
| 53 |
+
"41": "credit_limit",
|
| 54 |
+
"42": "oos",
|
| 55 |
+
"43": "shopping_list",
|
| 56 |
+
"44": "expiration_date",
|
| 57 |
+
"45": "routing",
|
| 58 |
+
"46": "meal_suggestion",
|
| 59 |
+
"47": "tire_change",
|
| 60 |
+
"48": "todo_list",
|
| 61 |
+
"49": "card_declined",
|
| 62 |
+
"50": "rewards_balance",
|
| 63 |
+
"51": "change_accent",
|
| 64 |
+
"52": "vaccines",
|
| 65 |
+
"53": "reminder_update",
|
| 66 |
+
"54": "food_last",
|
| 67 |
+
"55": "change_ai_name",
|
| 68 |
+
"56": "bill_due",
|
| 69 |
+
"57": "who_do_you_work_for",
|
| 70 |
+
"58": "share_location",
|
| 71 |
+
"59": "international_visa",
|
| 72 |
+
"60": "calendar",
|
| 73 |
+
"61": "translate",
|
| 74 |
+
"62": "carry_on",
|
| 75 |
+
"63": "book_flight",
|
| 76 |
+
"64": "insurance_change",
|
| 77 |
+
"65": "todo_list_update",
|
| 78 |
+
"66": "timezone",
|
| 79 |
+
"67": "cancel_reservation",
|
| 80 |
+
"68": "transactions",
|
| 81 |
+
"69": "credit_score",
|
| 82 |
+
"70": "report_fraud",
|
| 83 |
+
"71": "spending_history",
|
| 84 |
+
"72": "directions",
|
| 85 |
+
"73": "spelling",
|
| 86 |
+
"74": "insurance",
|
| 87 |
+
"75": "what_is_your_name",
|
| 88 |
+
"76": "reminder",
|
| 89 |
+
"77": "where_are_you_from",
|
| 90 |
+
"78": "distance",
|
| 91 |
+
"79": "payday",
|
| 92 |
+
"80": "flight_status",
|
| 93 |
+
"81": "find_phone",
|
| 94 |
+
"82": "greeting",
|
| 95 |
+
"83": "alarm",
|
| 96 |
+
"84": "order_status",
|
| 97 |
+
"85": "confirm_reservation",
|
| 98 |
+
"86": "cook_time",
|
| 99 |
+
"87": "damaged_card",
|
| 100 |
+
"88": "reset_settings",
|
| 101 |
+
"89": "pin_change",
|
| 102 |
+
"90": "replacement_card_duration",
|
| 103 |
+
"91": "new_card",
|
| 104 |
+
"92": "roll_dice",
|
| 105 |
+
"93": "income",
|
| 106 |
+
"94": "taxes",
|
| 107 |
+
"95": "date",
|
| 108 |
+
"96": "who_made_you",
|
| 109 |
+
"97": "pto_request",
|
| 110 |
+
"98": "tire_pressure",
|
| 111 |
+
"99": "how_old_are_you",
|
| 112 |
+
"100": "rollover_401k",
|
| 113 |
+
"101": "pto_request_status",
|
| 114 |
+
"102": "how_busy",
|
| 115 |
+
"103": "application_status",
|
| 116 |
+
"104": "recipe",
|
| 117 |
+
"105": "calendar_update",
|
| 118 |
+
"106": "play_music",
|
| 119 |
+
"107": "yes",
|
| 120 |
+
"108": "direct_deposit",
|
| 121 |
+
"109": "credit_limit_change",
|
| 122 |
+
"110": "gas",
|
| 123 |
+
"111": "pay_bill",
|
| 124 |
+
"112": "ingredients_list",
|
| 125 |
+
"113": "lost_luggage",
|
| 126 |
+
"114": "goodbye",
|
| 127 |
+
"115": "what_can_i_ask_you",
|
| 128 |
+
"116": "book_hotel",
|
| 129 |
+
"117": "are_you_a_bot",
|
| 130 |
+
"118": "next_song",
|
| 131 |
+
"119": "change_speed",
|
| 132 |
+
"120": "plug_type",
|
| 133 |
+
"121": "maybe",
|
| 134 |
+
"122": "w2",
|
| 135 |
+
"123": "oil_change_when",
|
| 136 |
+
"124": "thank_you",
|
| 137 |
+
"125": "shopping_list_update",
|
| 138 |
+
"126": "pto_balance",
|
| 139 |
+
"127": "order_checks",
|
| 140 |
+
"128": "travel_alert",
|
| 141 |
+
"129": "fun_fact",
|
| 142 |
+
"130": "sync_device",
|
| 143 |
+
"131": "schedule_maintenance",
|
| 144 |
+
"132": "apr",
|
| 145 |
+
"133": "transfer",
|
| 146 |
+
"134": "ingredient_substitution",
|
| 147 |
+
"135": "calories",
|
| 148 |
+
"136": "current_location",
|
| 149 |
+
"137": "international_fees",
|
| 150 |
+
"138": "calculator",
|
| 151 |
+
"139": "definition",
|
| 152 |
+
"140": "next_holiday",
|
| 153 |
+
"141": "update_playlist",
|
| 154 |
+
"142": "mpg",
|
| 155 |
+
"143": "min_payment",
|
| 156 |
+
"144": "change_user_name",
|
| 157 |
+
"145": "restaurant_suggestion",
|
| 158 |
+
"146": "travel_notification",
|
| 159 |
+
"147": "cancel",
|
| 160 |
+
"148": "pto_used",
|
| 161 |
+
"149": "travel_suggestion",
|
| 162 |
+
"150": "change_volume"
|
| 163 |
+
},
|
| 164 |
+
"initializer_range": 0.02,
|
| 165 |
+
"label2id": {
|
| 166 |
+
"accept_reservations": 9,
|
| 167 |
+
"account_blocked": 2,
|
| 168 |
+
"alarm": 83,
|
| 169 |
+
"application_status": 103,
|
| 170 |
+
"apr": 132,
|
| 171 |
+
"are_you_a_bot": 117,
|
| 172 |
+
"balance": 35,
|
| 173 |
+
"bill_balance": 27,
|
| 174 |
+
"bill_due": 56,
|
| 175 |
+
"book_flight": 63,
|
| 176 |
+
"book_hotel": 116,
|
| 177 |
+
"calculator": 138,
|
| 178 |
+
"calendar": 60,
|
| 179 |
+
"calendar_update": 105,
|
| 180 |
+
"calories": 135,
|
| 181 |
+
"cancel": 147,
|
| 182 |
+
"cancel_reservation": 67,
|
| 183 |
+
"car_rental": 40,
|
| 184 |
+
"card_declined": 49,
|
| 185 |
+
"carry_on": 62,
|
| 186 |
+
"change_accent": 51,
|
| 187 |
+
"change_ai_name": 55,
|
| 188 |
+
"change_language": 29,
|
| 189 |
+
"change_speed": 119,
|
| 190 |
+
"change_user_name": 144,
|
| 191 |
+
"change_volume": 150,
|
| 192 |
+
"confirm_reservation": 85,
|
| 193 |
+
"cook_time": 86,
|
| 194 |
+
"credit_limit": 41,
|
| 195 |
+
"credit_limit_change": 109,
|
| 196 |
+
"credit_score": 69,
|
| 197 |
+
"current_location": 136,
|
| 198 |
+
"damaged_card": 87,
|
| 199 |
+
"date": 95,
|
| 200 |
+
"definition": 139,
|
| 201 |
+
"direct_deposit": 108,
|
| 202 |
+
"directions": 72,
|
| 203 |
+
"distance": 78,
|
| 204 |
+
"do_you_have_pets": 34,
|
| 205 |
+
"exchange_rate": 38,
|
| 206 |
+
"expiration_date": 44,
|
| 207 |
+
"find_phone": 81,
|
| 208 |
+
"flight_status": 80,
|
| 209 |
+
"flip_coin": 33,
|
| 210 |
+
"food_last": 54,
|
| 211 |
+
"freeze_account": 20,
|
| 212 |
+
"fun_fact": 129,
|
| 213 |
+
"gas": 110,
|
| 214 |
+
"gas_type": 8,
|
| 215 |
+
"goodbye": 114,
|
| 216 |
+
"greeting": 82,
|
| 217 |
+
"how_busy": 102,
|
| 218 |
+
"how_old_are_you": 99,
|
| 219 |
+
"improve_credit_score": 28,
|
| 220 |
+
"income": 93,
|
| 221 |
+
"ingredient_substitution": 134,
|
| 222 |
+
"ingredients_list": 112,
|
| 223 |
+
"insurance": 74,
|
| 224 |
+
"insurance_change": 64,
|
| 225 |
+
"interest_rate": 7,
|
| 226 |
+
"international_fees": 137,
|
| 227 |
+
"international_visa": 59,
|
| 228 |
+
"jump_start": 17,
|
| 229 |
+
"last_maintenance": 37,
|
| 230 |
+
"lost_luggage": 113,
|
| 231 |
+
"make_call": 25,
|
| 232 |
+
"maybe": 121,
|
| 233 |
+
"meal_suggestion": 46,
|
| 234 |
+
"meaning_of_life": 22,
|
| 235 |
+
"measurement_conversion": 31,
|
| 236 |
+
"meeting_schedule": 19,
|
| 237 |
+
"min_payment": 143,
|
| 238 |
+
"mpg": 142,
|
| 239 |
+
"new_card": 91,
|
| 240 |
+
"next_holiday": 140,
|
| 241 |
+
"next_song": 118,
|
| 242 |
+
"no": 30,
|
| 243 |
+
"nutrition_info": 1,
|
| 244 |
+
"oil_change_how": 3,
|
| 245 |
+
"oil_change_when": 123,
|
| 246 |
+
"oos": 42,
|
| 247 |
+
"order": 16,
|
| 248 |
+
"order_checks": 127,
|
| 249 |
+
"order_status": 84,
|
| 250 |
+
"pay_bill": 111,
|
| 251 |
+
"payday": 79,
|
| 252 |
+
"pin_change": 89,
|
| 253 |
+
"play_music": 106,
|
| 254 |
+
"plug_type": 120,
|
| 255 |
+
"pto_balance": 126,
|
| 256 |
+
"pto_request": 97,
|
| 257 |
+
"pto_request_status": 101,
|
| 258 |
+
"pto_used": 148,
|
| 259 |
+
"recipe": 104,
|
| 260 |
+
"redeem_rewards": 6,
|
| 261 |
+
"reminder": 76,
|
| 262 |
+
"reminder_update": 53,
|
| 263 |
+
"repeat": 13,
|
| 264 |
+
"replacement_card_duration": 90,
|
| 265 |
+
"report_fraud": 70,
|
| 266 |
+
"report_lost_card": 12,
|
| 267 |
+
"reset_settings": 88,
|
| 268 |
+
"restaurant_reservation": 23,
|
| 269 |
+
"restaurant_reviews": 0,
|
| 270 |
+
"restaurant_suggestion": 145,
|
| 271 |
+
"rewards_balance": 50,
|
| 272 |
+
"roll_dice": 92,
|
| 273 |
+
"rollover_401k": 100,
|
| 274 |
+
"routing": 45,
|
| 275 |
+
"schedule_maintenance": 131,
|
| 276 |
+
"schedule_meeting": 18,
|
| 277 |
+
"share_location": 58,
|
| 278 |
+
"shopping_list": 43,
|
| 279 |
+
"shopping_list_update": 125,
|
| 280 |
+
"smart_home": 10,
|
| 281 |
+
"spelling": 73,
|
| 282 |
+
"spending_history": 71,
|
| 283 |
+
"sync_device": 130,
|
| 284 |
+
"taxes": 94,
|
| 285 |
+
"tell_joke": 36,
|
| 286 |
+
"text": 26,
|
| 287 |
+
"thank_you": 124,
|
| 288 |
+
"time": 4,
|
| 289 |
+
"timer": 32,
|
| 290 |
+
"timezone": 66,
|
| 291 |
+
"tire_change": 47,
|
| 292 |
+
"tire_pressure": 98,
|
| 293 |
+
"todo_list": 48,
|
| 294 |
+
"todo_list_update": 65,
|
| 295 |
+
"traffic": 24,
|
| 296 |
+
"transactions": 68,
|
| 297 |
+
"transfer": 133,
|
| 298 |
+
"translate": 61,
|
| 299 |
+
"travel_alert": 128,
|
| 300 |
+
"travel_notification": 146,
|
| 301 |
+
"travel_suggestion": 149,
|
| 302 |
+
"uber": 39,
|
| 303 |
+
"update_playlist": 141,
|
| 304 |
+
"user_name": 11,
|
| 305 |
+
"vaccines": 52,
|
| 306 |
+
"w2": 122,
|
| 307 |
+
"weather": 5,
|
| 308 |
+
"what_are_your_hobbies": 15,
|
| 309 |
+
"what_can_i_ask_you": 115,
|
| 310 |
+
"what_is_your_name": 75,
|
| 311 |
+
"what_song": 21,
|
| 312 |
+
"where_are_you_from": 77,
|
| 313 |
+
"whisper_mode": 14,
|
| 314 |
+
"who_do_you_work_for": 57,
|
| 315 |
+
"who_made_you": 96,
|
| 316 |
+
"yes": 107
|
| 317 |
+
},
|
| 318 |
+
"max_position_embeddings": 512,
|
| 319 |
+
"model_type": "distilbert",
|
| 320 |
+
"n_heads": 12,
|
| 321 |
+
"n_layers": 6,
|
| 322 |
+
"pad_token_id": 0,
|
| 323 |
+
"problem_type": "single_label_classification",
|
| 324 |
+
"qa_dropout": 0.1,
|
| 325 |
+
"seq_classif_dropout": 0.2,
|
| 326 |
+
"sinusoidal_pos_embds": false,
|
| 327 |
+
"tie_weights_": true,
|
| 328 |
+
"torch_dtype": "float32",
|
| 329 |
+
"transformers_version": "4.42.4",
|
| 330 |
+
"vocab_size": 30522
|
| 331 |
+
}
|
run-8/checkpoint-1000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2265efdbd127f32817bf80110d8d6f541a55048b2e22b7e0e6227a688a9ef047
|
| 3 |
+
size 268290900
|
run-8/checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4557295dc599e62a6393fbd116ab5c4958396bbe86258e055f7094ff579a4f6
|
| 3 |
+
size 536643898
|
run-8/checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d11e6238fdff466b11616d085da1d3114fa87946b89c0aaea7f53ce2f94103c9
|
| 3 |
+
size 14244
|
run-8/checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:72a4307cca951ef6d42c7f184774e8b1259c78b2bf130b8443ec0b62f6dbf689
|
| 3 |
+
size 1064
|
run-8/checkpoint-1000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
run-8/checkpoint-1000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
run-8/checkpoint-1000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "DistilBertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
run-8/checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.1446540880503147,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1000,
|
| 7 |
+
"is_hyper_param_search": true,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.5887096774193549,
|
| 14 |
+
"eval_loss": 0.19886387884616852,
|
| 15 |
+
"eval_runtime": 5.5771,
|
| 16 |
+
"eval_samples_per_second": 555.842,
|
| 17 |
+
"eval_steps_per_second": 11.655,
|
| 18 |
+
"step": 318
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 1.5723270440251573,
|
| 22 |
+
"grad_norm": 0.5294517874717712,
|
| 23 |
+
"learning_rate": 1.650593990216632e-05,
|
| 24 |
+
"loss": 0.3203,
|
| 25 |
+
"step": 500
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 2.0,
|
| 29 |
+
"eval_accuracy": 0.8193548387096774,
|
| 30 |
+
"eval_loss": 0.09618715196847916,
|
| 31 |
+
"eval_runtime": 5.9897,
|
| 32 |
+
"eval_samples_per_second": 517.551,
|
| 33 |
+
"eval_steps_per_second": 10.852,
|
| 34 |
+
"step": 636
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.8783870967741936,
|
| 39 |
+
"eval_loss": 0.06421981006860733,
|
| 40 |
+
"eval_runtime": 5.7112,
|
| 41 |
+
"eval_samples_per_second": 542.789,
|
| 42 |
+
"eval_steps_per_second": 11.381,
|
| 43 |
+
"step": 954
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.1446540880503147,
|
| 47 |
+
"grad_norm": 0.44660821557044983,
|
| 48 |
+
"learning_rate": 1.3011879804332637e-05,
|
| 49 |
+
"loss": 0.1116,
|
| 50 |
+
"step": 1000
|
| 51 |
+
}
|
| 52 |
+
],
|
| 53 |
+
"logging_steps": 500,
|
| 54 |
+
"max_steps": 2862,
|
| 55 |
+
"num_input_tokens_seen": 0,
|
| 56 |
+
"num_train_epochs": 9,
|
| 57 |
+
"save_steps": 500,
|
| 58 |
+
"stateful_callbacks": {
|
| 59 |
+
"TrainerControl": {
|
| 60 |
+
"args": {
|
| 61 |
+
"should_epoch_stop": false,
|
| 62 |
+
"should_evaluate": false,
|
| 63 |
+
"should_log": false,
|
| 64 |
+
"should_save": true,
|
| 65 |
+
"should_training_stop": false
|
| 66 |
+
},
|
| 67 |
+
"attributes": {}
|
| 68 |
+
}
|
| 69 |
+
},
|
| 70 |
+
"total_flos": 260941334653608.0,
|
| 71 |
+
"train_batch_size": 48,
|
| 72 |
+
"trial_name": null,
|
| 73 |
+
"trial_params": {
|
| 74 |
+
"alpha": 0.9438886679483393,
|
| 75 |
+
"num_train_epochs": 9,
|
| 76 |
+
"temperature": 11
|
| 77 |
+
}
|
| 78 |
+
}
|
run-8/checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba03b00bb75688886f2bbe651ac208f30c9d535b120c047c2f2dc251bf5d4097
|
| 3 |
+
size 5176
|
run-8/checkpoint-1000/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
run-8/checkpoint-500/config.json
ADDED
|
@@ -0,0 +1,331 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert-base-uncased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DistilBertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "restaurant_reviews",
|
| 13 |
+
"1": "nutrition_info",
|
| 14 |
+
"2": "account_blocked",
|
| 15 |
+
"3": "oil_change_how",
|
| 16 |
+
"4": "time",
|
| 17 |
+
"5": "weather",
|
| 18 |
+
"6": "redeem_rewards",
|
| 19 |
+
"7": "interest_rate",
|
| 20 |
+
"8": "gas_type",
|
| 21 |
+
"9": "accept_reservations",
|
| 22 |
+
"10": "smart_home",
|
| 23 |
+
"11": "user_name",
|
| 24 |
+
"12": "report_lost_card",
|
| 25 |
+
"13": "repeat",
|
| 26 |
+
"14": "whisper_mode",
|
| 27 |
+
"15": "what_are_your_hobbies",
|
| 28 |
+
"16": "order",
|
| 29 |
+
"17": "jump_start",
|
| 30 |
+
"18": "schedule_meeting",
|
| 31 |
+
"19": "meeting_schedule",
|
| 32 |
+
"20": "freeze_account",
|
| 33 |
+
"21": "what_song",
|
| 34 |
+
"22": "meaning_of_life",
|
| 35 |
+
"23": "restaurant_reservation",
|
| 36 |
+
"24": "traffic",
|
| 37 |
+
"25": "make_call",
|
| 38 |
+
"26": "text",
|
| 39 |
+
"27": "bill_balance",
|
| 40 |
+
"28": "improve_credit_score",
|
| 41 |
+
"29": "change_language",
|
| 42 |
+
"30": "no",
|
| 43 |
+
"31": "measurement_conversion",
|
| 44 |
+
"32": "timer",
|
| 45 |
+
"33": "flip_coin",
|
| 46 |
+
"34": "do_you_have_pets",
|
| 47 |
+
"35": "balance",
|
| 48 |
+
"36": "tell_joke",
|
| 49 |
+
"37": "last_maintenance",
|
| 50 |
+
"38": "exchange_rate",
|
| 51 |
+
"39": "uber",
|
| 52 |
+
"40": "car_rental",
|
| 53 |
+
"41": "credit_limit",
|
| 54 |
+
"42": "oos",
|
| 55 |
+
"43": "shopping_list",
|
| 56 |
+
"44": "expiration_date",
|
| 57 |
+
"45": "routing",
|
| 58 |
+
"46": "meal_suggestion",
|
| 59 |
+
"47": "tire_change",
|
| 60 |
+
"48": "todo_list",
|
| 61 |
+
"49": "card_declined",
|
| 62 |
+
"50": "rewards_balance",
|
| 63 |
+
"51": "change_accent",
|
| 64 |
+
"52": "vaccines",
|
| 65 |
+
"53": "reminder_update",
|
| 66 |
+
"54": "food_last",
|
| 67 |
+
"55": "change_ai_name",
|
| 68 |
+
"56": "bill_due",
|
| 69 |
+
"57": "who_do_you_work_for",
|
| 70 |
+
"58": "share_location",
|
| 71 |
+
"59": "international_visa",
|
| 72 |
+
"60": "calendar",
|
| 73 |
+
"61": "translate",
|
| 74 |
+
"62": "carry_on",
|
| 75 |
+
"63": "book_flight",
|
| 76 |
+
"64": "insurance_change",
|
| 77 |
+
"65": "todo_list_update",
|
| 78 |
+
"66": "timezone",
|
| 79 |
+
"67": "cancel_reservation",
|
| 80 |
+
"68": "transactions",
|
| 81 |
+
"69": "credit_score",
|
| 82 |
+
"70": "report_fraud",
|
| 83 |
+
"71": "spending_history",
|
| 84 |
+
"72": "directions",
|
| 85 |
+
"73": "spelling",
|
| 86 |
+
"74": "insurance",
|
| 87 |
+
"75": "what_is_your_name",
|
| 88 |
+
"76": "reminder",
|
| 89 |
+
"77": "where_are_you_from",
|
| 90 |
+
"78": "distance",
|
| 91 |
+
"79": "payday",
|
| 92 |
+
"80": "flight_status",
|
| 93 |
+
"81": "find_phone",
|
| 94 |
+
"82": "greeting",
|
| 95 |
+
"83": "alarm",
|
| 96 |
+
"84": "order_status",
|
| 97 |
+
"85": "confirm_reservation",
|
| 98 |
+
"86": "cook_time",
|
| 99 |
+
"87": "damaged_card",
|
| 100 |
+
"88": "reset_settings",
|
| 101 |
+
"89": "pin_change",
|
| 102 |
+
"90": "replacement_card_duration",
|
| 103 |
+
"91": "new_card",
|
| 104 |
+
"92": "roll_dice",
|
| 105 |
+
"93": "income",
|
| 106 |
+
"94": "taxes",
|
| 107 |
+
"95": "date",
|
| 108 |
+
"96": "who_made_you",
|
| 109 |
+
"97": "pto_request",
|
| 110 |
+
"98": "tire_pressure",
|
| 111 |
+
"99": "how_old_are_you",
|
| 112 |
+
"100": "rollover_401k",
|
| 113 |
+
"101": "pto_request_status",
|
| 114 |
+
"102": "how_busy",
|
| 115 |
+
"103": "application_status",
|
| 116 |
+
"104": "recipe",
|
| 117 |
+
"105": "calendar_update",
|
| 118 |
+
"106": "play_music",
|
| 119 |
+
"107": "yes",
|
| 120 |
+
"108": "direct_deposit",
|
| 121 |
+
"109": "credit_limit_change",
|
| 122 |
+
"110": "gas",
|
| 123 |
+
"111": "pay_bill",
|
| 124 |
+
"112": "ingredients_list",
|
| 125 |
+
"113": "lost_luggage",
|
| 126 |
+
"114": "goodbye",
|
| 127 |
+
"115": "what_can_i_ask_you",
|
| 128 |
+
"116": "book_hotel",
|
| 129 |
+
"117": "are_you_a_bot",
|
| 130 |
+
"118": "next_song",
|
| 131 |
+
"119": "change_speed",
|
| 132 |
+
"120": "plug_type",
|
| 133 |
+
"121": "maybe",
|
| 134 |
+
"122": "w2",
|
| 135 |
+
"123": "oil_change_when",
|
| 136 |
+
"124": "thank_you",
|
| 137 |
+
"125": "shopping_list_update",
|
| 138 |
+
"126": "pto_balance",
|
| 139 |
+
"127": "order_checks",
|
| 140 |
+
"128": "travel_alert",
|
| 141 |
+
"129": "fun_fact",
|
| 142 |
+
"130": "sync_device",
|
| 143 |
+
"131": "schedule_maintenance",
|
| 144 |
+
"132": "apr",
|
| 145 |
+
"133": "transfer",
|
| 146 |
+
"134": "ingredient_substitution",
|
| 147 |
+
"135": "calories",
|
| 148 |
+
"136": "current_location",
|
| 149 |
+
"137": "international_fees",
|
| 150 |
+
"138": "calculator",
|
| 151 |
+
"139": "definition",
|
| 152 |
+
"140": "next_holiday",
|
| 153 |
+
"141": "update_playlist",
|
| 154 |
+
"142": "mpg",
|
| 155 |
+
"143": "min_payment",
|
| 156 |
+
"144": "change_user_name",
|
| 157 |
+
"145": "restaurant_suggestion",
|
| 158 |
+
"146": "travel_notification",
|
| 159 |
+
"147": "cancel",
|
| 160 |
+
"148": "pto_used",
|
| 161 |
+
"149": "travel_suggestion",
|
| 162 |
+
"150": "change_volume"
|
| 163 |
+
},
|
| 164 |
+
"initializer_range": 0.02,
|
| 165 |
+
"label2id": {
|
| 166 |
+
"accept_reservations": 9,
|
| 167 |
+
"account_blocked": 2,
|
| 168 |
+
"alarm": 83,
|
| 169 |
+
"application_status": 103,
|
| 170 |
+
"apr": 132,
|
| 171 |
+
"are_you_a_bot": 117,
|
| 172 |
+
"balance": 35,
|
| 173 |
+
"bill_balance": 27,
|
| 174 |
+
"bill_due": 56,
|
| 175 |
+
"book_flight": 63,
|
| 176 |
+
"book_hotel": 116,
|
| 177 |
+
"calculator": 138,
|
| 178 |
+
"calendar": 60,
|
| 179 |
+
"calendar_update": 105,
|
| 180 |
+
"calories": 135,
|
| 181 |
+
"cancel": 147,
|
| 182 |
+
"cancel_reservation": 67,
|
| 183 |
+
"car_rental": 40,
|
| 184 |
+
"card_declined": 49,
|
| 185 |
+
"carry_on": 62,
|
| 186 |
+
"change_accent": 51,
|
| 187 |
+
"change_ai_name": 55,
|
| 188 |
+
"change_language": 29,
|
| 189 |
+
"change_speed": 119,
|
| 190 |
+
"change_user_name": 144,
|
| 191 |
+
"change_volume": 150,
|
| 192 |
+
"confirm_reservation": 85,
|
| 193 |
+
"cook_time": 86,
|
| 194 |
+
"credit_limit": 41,
|
| 195 |
+
"credit_limit_change": 109,
|
| 196 |
+
"credit_score": 69,
|
| 197 |
+
"current_location": 136,
|
| 198 |
+
"damaged_card": 87,
|
| 199 |
+
"date": 95,
|
| 200 |
+
"definition": 139,
|
| 201 |
+
"direct_deposit": 108,
|
| 202 |
+
"directions": 72,
|
| 203 |
+
"distance": 78,
|
| 204 |
+
"do_you_have_pets": 34,
|
| 205 |
+
"exchange_rate": 38,
|
| 206 |
+
"expiration_date": 44,
|
| 207 |
+
"find_phone": 81,
|
| 208 |
+
"flight_status": 80,
|
| 209 |
+
"flip_coin": 33,
|
| 210 |
+
"food_last": 54,
|
| 211 |
+
"freeze_account": 20,
|
| 212 |
+
"fun_fact": 129,
|
| 213 |
+
"gas": 110,
|
| 214 |
+
"gas_type": 8,
|
| 215 |
+
"goodbye": 114,
|
| 216 |
+
"greeting": 82,
|
| 217 |
+
"how_busy": 102,
|
| 218 |
+
"how_old_are_you": 99,
|
| 219 |
+
"improve_credit_score": 28,
|
| 220 |
+
"income": 93,
|
| 221 |
+
"ingredient_substitution": 134,
|
| 222 |
+
"ingredients_list": 112,
|
| 223 |
+
"insurance": 74,
|
| 224 |
+
"insurance_change": 64,
|
| 225 |
+
"interest_rate": 7,
|
| 226 |
+
"international_fees": 137,
|
| 227 |
+
"international_visa": 59,
|
| 228 |
+
"jump_start": 17,
|
| 229 |
+
"last_maintenance": 37,
|
| 230 |
+
"lost_luggage": 113,
|
| 231 |
+
"make_call": 25,
|
| 232 |
+
"maybe": 121,
|
| 233 |
+
"meal_suggestion": 46,
|
| 234 |
+
"meaning_of_life": 22,
|
| 235 |
+
"measurement_conversion": 31,
|
| 236 |
+
"meeting_schedule": 19,
|
| 237 |
+
"min_payment": 143,
|
| 238 |
+
"mpg": 142,
|
| 239 |
+
"new_card": 91,
|
| 240 |
+
"next_holiday": 140,
|
| 241 |
+
"next_song": 118,
|
| 242 |
+
"no": 30,
|
| 243 |
+
"nutrition_info": 1,
|
| 244 |
+
"oil_change_how": 3,
|
| 245 |
+
"oil_change_when": 123,
|
| 246 |
+
"oos": 42,
|
| 247 |
+
"order": 16,
|
| 248 |
+
"order_checks": 127,
|
| 249 |
+
"order_status": 84,
|
| 250 |
+
"pay_bill": 111,
|
| 251 |
+
"payday": 79,
|
| 252 |
+
"pin_change": 89,
|
| 253 |
+
"play_music": 106,
|
| 254 |
+
"plug_type": 120,
|
| 255 |
+
"pto_balance": 126,
|
| 256 |
+
"pto_request": 97,
|
| 257 |
+
"pto_request_status": 101,
|
| 258 |
+
"pto_used": 148,
|
| 259 |
+
"recipe": 104,
|
| 260 |
+
"redeem_rewards": 6,
|
| 261 |
+
"reminder": 76,
|
| 262 |
+
"reminder_update": 53,
|
| 263 |
+
"repeat": 13,
|
| 264 |
+
"replacement_card_duration": 90,
|
| 265 |
+
"report_fraud": 70,
|
| 266 |
+
"report_lost_card": 12,
|
| 267 |
+
"reset_settings": 88,
|
| 268 |
+
"restaurant_reservation": 23,
|
| 269 |
+
"restaurant_reviews": 0,
|
| 270 |
+
"restaurant_suggestion": 145,
|
| 271 |
+
"rewards_balance": 50,
|
| 272 |
+
"roll_dice": 92,
|
| 273 |
+
"rollover_401k": 100,
|
| 274 |
+
"routing": 45,
|
| 275 |
+
"schedule_maintenance": 131,
|
| 276 |
+
"schedule_meeting": 18,
|
| 277 |
+
"share_location": 58,
|
| 278 |
+
"shopping_list": 43,
|
| 279 |
+
"shopping_list_update": 125,
|
| 280 |
+
"smart_home": 10,
|
| 281 |
+
"spelling": 73,
|
| 282 |
+
"spending_history": 71,
|
| 283 |
+
"sync_device": 130,
|
| 284 |
+
"taxes": 94,
|
| 285 |
+
"tell_joke": 36,
|
| 286 |
+
"text": 26,
|
| 287 |
+
"thank_you": 124,
|
| 288 |
+
"time": 4,
|
| 289 |
+
"timer": 32,
|
| 290 |
+
"timezone": 66,
|
| 291 |
+
"tire_change": 47,
|
| 292 |
+
"tire_pressure": 98,
|
| 293 |
+
"todo_list": 48,
|
| 294 |
+
"todo_list_update": 65,
|
| 295 |
+
"traffic": 24,
|
| 296 |
+
"transactions": 68,
|
| 297 |
+
"transfer": 133,
|
| 298 |
+
"translate": 61,
|
| 299 |
+
"travel_alert": 128,
|
| 300 |
+
"travel_notification": 146,
|
| 301 |
+
"travel_suggestion": 149,
|
| 302 |
+
"uber": 39,
|
| 303 |
+
"update_playlist": 141,
|
| 304 |
+
"user_name": 11,
|
| 305 |
+
"vaccines": 52,
|
| 306 |
+
"w2": 122,
|
| 307 |
+
"weather": 5,
|
| 308 |
+
"what_are_your_hobbies": 15,
|
| 309 |
+
"what_can_i_ask_you": 115,
|
| 310 |
+
"what_is_your_name": 75,
|
| 311 |
+
"what_song": 21,
|
| 312 |
+
"where_are_you_from": 77,
|
| 313 |
+
"whisper_mode": 14,
|
| 314 |
+
"who_do_you_work_for": 57,
|
| 315 |
+
"who_made_you": 96,
|
| 316 |
+
"yes": 107
|
| 317 |
+
},
|
| 318 |
+
"max_position_embeddings": 512,
|
| 319 |
+
"model_type": "distilbert",
|
| 320 |
+
"n_heads": 12,
|
| 321 |
+
"n_layers": 6,
|
| 322 |
+
"pad_token_id": 0,
|
| 323 |
+
"problem_type": "single_label_classification",
|
| 324 |
+
"qa_dropout": 0.1,
|
| 325 |
+
"seq_classif_dropout": 0.2,
|
| 326 |
+
"sinusoidal_pos_embds": false,
|
| 327 |
+
"tie_weights_": true,
|
| 328 |
+
"torch_dtype": "float32",
|
| 329 |
+
"transformers_version": "4.42.4",
|
| 330 |
+
"vocab_size": 30522
|
| 331 |
+
}
|
run-8/checkpoint-500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33620587b37c196969d3f412f74ef041dcfeee8258db5933dc3c4540e70b1e43
|
| 3 |
+
size 268290900
|
run-8/checkpoint-500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3bab979622485a1e70f82a64c3ec4aae99b0262391ec447afad3b7eae33df5ce
|
| 3 |
+
size 536643898
|
run-8/checkpoint-500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9b52f07ff6d370f2e4d5846adc92d047fdffc95329c677c7e7ee44327b91aef
|
| 3 |
+
size 14244
|