Upload folder using huggingface_hub
Browse files- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +41 -3
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4991044576
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3eacff62a23c0f57c841d757c3de0176f8e05b70adb5fef1b0d6d80d8dd225f
|
| 3 |
size 4991044576
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79d89dee4b4f65f8d68127ad11dd7780f1cbe7d89d9f328c68349782405e00e5
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4588398592
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e0665614c8ff1566c179a967d1d2b32dcfff7fed86afb6bb159a6e963381207
|
| 3 |
size 4588398592
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 29158711534
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7bfb2dcf720508944379eaa7b6e959c1f7b90ff992889e28a26237581598e99b
|
| 3 |
size 29158711534
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9f6ecf2f8e43e1a548a602a630d683375015193ff9ea4b05cc043ce629aadad
|
| 3 |
size 14244
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bcbbe9f35b4d475722d8a6ee3bd3440cc04fe40aa519263e8ff51d152a5551d8
|
| 3 |
size 1000
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 1.
|
| 5 |
"eval_steps": 93,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -811,6 +811,44 @@
|
|
| 811 |
"eval_validation_altalexprivacy_samples_per_second": 1.355,
|
| 812 |
"eval_validation_altalexprivacy_steps_per_second": 0.677,
|
| 813 |
"step": 1953
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 814 |
}
|
| 815 |
],
|
| 816 |
"logging_steps": 93,
|
|
@@ -818,7 +856,7 @@
|
|
| 818 |
"num_input_tokens_seen": 0,
|
| 819 |
"num_train_epochs": 2,
|
| 820 |
"save_steps": 93,
|
| 821 |
-
"total_flos": 2.
|
| 822 |
"train_batch_size": 1,
|
| 823 |
"trial_name": null,
|
| 824 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.7061552092562673,
|
| 5 |
"eval_steps": 93,
|
| 6 |
+
"global_step": 2046,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 811 |
"eval_validation_altalexprivacy_samples_per_second": 1.355,
|
| 812 |
"eval_validation_altalexprivacy_steps_per_second": 0.677,
|
| 813 |
"step": 1953
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 1.71,
|
| 817 |
+
"learning_rate": 4.000000000000001e-06,
|
| 818 |
+
"loss": 1.1672,
|
| 819 |
+
"step": 2046
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 1.71,
|
| 823 |
+
"eval_validation_privacy_sources_loss": 0.9668792486190796,
|
| 824 |
+
"eval_validation_privacy_sources_runtime": 169.8267,
|
| 825 |
+
"eval_validation_privacy_sources_samples_per_second": 1.354,
|
| 826 |
+
"eval_validation_privacy_sources_steps_per_second": 0.677,
|
| 827 |
+
"step": 2046
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 1.71,
|
| 831 |
+
"eval_validation_agenda_digitale_loss": 1.6615241765975952,
|
| 832 |
+
"eval_validation_agenda_digitale_runtime": 160.9234,
|
| 833 |
+
"eval_validation_agenda_digitale_samples_per_second": 1.355,
|
| 834 |
+
"eval_validation_agenda_digitale_steps_per_second": 0.677,
|
| 835 |
+
"step": 2046
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.71,
|
| 839 |
+
"eval_validation_leggepertutti_loss": 1.5065767765045166,
|
| 840 |
+
"eval_validation_leggepertutti_runtime": 28.0518,
|
| 841 |
+
"eval_validation_leggepertutti_samples_per_second": 1.355,
|
| 842 |
+
"eval_validation_leggepertutti_steps_per_second": 0.677,
|
| 843 |
+
"step": 2046
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 1.71,
|
| 847 |
+
"eval_validation_altalexprivacy_loss": 1.464535117149353,
|
| 848 |
+
"eval_validation_altalexprivacy_runtime": 45.7867,
|
| 849 |
+
"eval_validation_altalexprivacy_samples_per_second": 1.354,
|
| 850 |
+
"eval_validation_altalexprivacy_steps_per_second": 0.677,
|
| 851 |
+
"step": 2046
|
| 852 |
}
|
| 853 |
],
|
| 854 |
"logging_steps": 93,
|
|
|
|
| 856 |
"num_input_tokens_seen": 0,
|
| 857 |
"num_train_epochs": 2,
|
| 858 |
"save_steps": 93,
|
| 859 |
+
"total_flos": 2.295971565472737e+19,
|
| 860 |
"train_batch_size": 1,
|
| 861 |
"trial_name": null,
|
| 862 |
"trial_params": null
|