Hugging Face's logo

MarkelFe
/
PoliticalSpeech2

Text Generation
Transformers
PyTorch
gpt2
text-generation-inference
Model card Files
xet
MarkelFe commited on
Commit
0a7110f
·
1 Parent(s): 1bd7110

Training in progress, step 310000

Browse files
Files changed (6) hide show
  1. last-checkpoint/optimizer.pt +1 -1
  2. last-checkpoint/pytorch_model.bin +1 -1
  3. last-checkpoint/rng_state.pth +1 -1
  4. last-checkpoint/scheduler.pt +1 -1
  5. last-checkpoint/trainer_state.json +131 -3
  6. pytorch_model.bin +1 -1
last-checkpoint/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6e9dfb9a273258a4da0cd0abf9c473733d64de15c9b436a4e310a7516992aadc
3
  size 995605445
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:834b3f1780c9fd89f6bfa5f319b65a6486e6af60221f901d7e860780ad39830b
3
  size 995605445
last-checkpoint/pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f604aecb41e9633c05e8dd317b656732e5079082b765a72f0a34dfa3c306d7d0
3
  size 510398013
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b8e5d8ebe593ce6ffea74db7aefd5f8a2df25512838f857c42fe859f241094e
3
  size 510398013
last-checkpoint/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:24a2cf6f432c6ef4028563f06584b1577c26f344c8073d1e42dc72f657ef2033
3
  size 14575
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77c94379f004ea0bd3b6c24318506613489998b98c0479ee6ca058b4723d9413
3
  size 14575
last-checkpoint/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fb57dbd8526d420080443a568b84c7a184bdfc060adbe06867b56eb2f70fa448
3
  size 627
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea74371489dafdb6947d1051749433c783541685907f59e03a46eb374dc6453a
3
  size 627
last-checkpoint/trainer_state.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 7.102945354673738,
5
- "global_step": 300000,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
@@ -3852,11 +3852,139 @@
3852
  "eval_samples_per_second": 166.395,
3853
  "eval_steps_per_second": 20.804,
3854
  "step": 300000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3855
  }
3856
  ],
3857
  "max_steps": 633540,
3858
  "num_train_epochs": 15,
3859
- "total_flos": 6.3395167319424e+16,
3860
  "trial_name": null,
3861
  "trial_params": null
3862
  }
 
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 7.339710199829529,
5
+ "global_step": 310000,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
 
3852
  "eval_samples_per_second": 166.395,
3853
  "eval_steps_per_second": 20.804,
3854
  "step": 300000
3855
+ },
3856
+ {
3857
+ "epoch": 7.11,
3858
+ "learning_rate": 1.1886226599741138e-06,
3859
+ "loss": 2.4182,
3860
+ "step": 300500
3861
+ },
3862
+ {
3863
+ "epoch": 7.13,
3864
+ "learning_rate": 1.1254853679325695e-06,
3865
+ "loss": 2.4228,
3866
+ "step": 301000
3867
+ },
3868
+ {
3869
+ "epoch": 7.14,
3870
+ "learning_rate": 1.062348075891025e-06,
3871
+ "loss": 2.4385,
3872
+ "step": 301500
3873
+ },
3874
+ {
3875
+ "epoch": 7.15,
3876
+ "learning_rate": 9.992107838494808e-07,
3877
+ "loss": 2.4361,
3878
+ "step": 302000
3879
+ },
3880
+ {
3881
+ "epoch": 7.16,
3882
+ "learning_rate": 9.360734918079364e-07,
3883
+ "loss": 2.43,
3884
+ "step": 302500
3885
+ },
3886
+ {
3887
+ "epoch": 7.17,
3888
+ "learning_rate": 8.729361997663921e-07,
3889
+ "loss": 2.4245,
3890
+ "step": 303000
3891
+ },
3892
+ {
3893
+ "epoch": 7.19,
3894
+ "learning_rate": 8.097989077248477e-07,
3895
+ "loss": 2.4383,
3896
+ "step": 303500
3897
+ },
3898
+ {
3899
+ "epoch": 7.2,
3900
+ "learning_rate": 7.466616156833035e-07,
3901
+ "loss": 2.4207,
3902
+ "step": 304000
3903
+ },
3904
+ {
3905
+ "epoch": 7.21,
3906
+ "learning_rate": 6.835243236417591e-07,
3907
+ "loss": 2.4286,
3908
+ "step": 304500
3909
+ },
3910
+ {
3911
+ "epoch": 7.22,
3912
+ "learning_rate": 6.203870316002147e-07,
3913
+ "loss": 2.4143,
3914
+ "step": 305000
3915
+ },
3916
+ {
3917
+ "epoch": 7.23,
3918
+ "learning_rate": 5.572497395586704e-07,
3919
+ "loss": 2.4354,
3920
+ "step": 305500
3921
+ },
3922
+ {
3923
+ "epoch": 7.25,
3924
+ "learning_rate": 4.94112447517126e-07,
3925
+ "loss": 2.4345,
3926
+ "step": 306000
3927
+ },
3928
+ {
3929
+ "epoch": 7.26,
3930
+ "learning_rate": 4.309751554755817e-07,
3931
+ "loss": 2.4294,
3932
+ "step": 306500
3933
+ },
3934
+ {
3935
+ "epoch": 7.27,
3936
+ "learning_rate": 3.678378634340374e-07,
3937
+ "loss": 2.4442,
3938
+ "step": 307000
3939
+ },
3940
+ {
3941
+ "epoch": 7.28,
3942
+ "learning_rate": 3.04700571392493e-07,
3943
+ "loss": 2.4416,
3944
+ "step": 307500
3945
+ },
3946
+ {
3947
+ "epoch": 7.29,
3948
+ "learning_rate": 2.4156327935094866e-07,
3949
+ "loss": 2.4323,
3950
+ "step": 308000
3951
+ },
3952
+ {
3953
+ "epoch": 7.3,
3954
+ "learning_rate": 1.7842598730940432e-07,
3955
+ "loss": 2.4233,
3956
+ "step": 308500
3957
+ },
3958
+ {
3959
+ "epoch": 7.32,
3960
+ "learning_rate": 1.1528869526785997e-07,
3961
+ "loss": 2.4391,
3962
+ "step": 309000
3963
+ },
3964
+ {
3965
+ "epoch": 7.33,
3966
+ "learning_rate": 5.2151403226315624e-08,
3967
+ "loss": 2.442,
3968
+ "step": 309500
3969
+ },
3970
+ {
3971
+ "epoch": 7.34,
3972
+ "learning_rate": 0.0,
3973
+ "loss": 2.4491,
3974
+ "step": 310000
3975
+ },
3976
+ {
3977
+ "epoch": 7.34,
3978
+ "eval_loss": 3.1522228717803955,
3979
+ "eval_runtime": 112.8215,
3980
+ "eval_samples_per_second": 166.387,
3981
+ "eval_steps_per_second": 20.803,
3982
+ "step": 310000
3983
  }
3984
  ],
3985
  "max_steps": 633540,
3986
  "num_train_epochs": 15,
3987
+ "total_flos": 6.5501895567744e+16,
3988
  "trial_name": null,
3989
  "trial_params": null
3990
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f604aecb41e9633c05e8dd317b656732e5079082b765a72f0a34dfa3c306d7d0
3
  size 510398013
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b8e5d8ebe593ce6ffea74db7aefd5f8a2df25512838f857c42fe859f241094e
3
  size 510398013