Training in progress, step 12000, checkpoint

last-checkpoint/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cd21524818a9658a40c8f2457f7769a1d67509b6807751bbbc95d8685aa23d53
+oid sha256:4d089788bfa72e7477a93f012139d08f9411e567869d6d9a8c9af060f6708360
 size 2141578

last-checkpoint/global_step12000/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23035022a18fee5ab2143537ab19b19f5e2ad1444b4fc313fa8c219673728646
+size 3369573916

last-checkpoint/global_step12000/zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5a1ec2b4a684624b9dcd34806bf8cb62c609d8afd7006755943a234f1650796
+size 3149677

last-checkpoint/global_step12000/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8990b2680e89da2aac191a3d88b3942ba0106ebd003859840cce1d69d64148dc
+size 3369573916

last-checkpoint/global_step12000/zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9ca228b6dc913363798a6e2055bfc902a0e2ad7a9eeef7a95934288d98f0d7c
+size 3149677

last-checkpoint/global_step12000/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7f9a7d067a18ebce018c895425fa60691fa184676416964f7fd3954d688001b
+size 3369573916

last-checkpoint/global_step12000/zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48491e1f7c2255f416115e6c5a4e1d8311c8d2cdf3186ab4d8e3e6e1cef44766
+size 3149677

last-checkpoint/global_step12000/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72f553d7a0e6fae7824abfb7355ddcce94f88c68372af98a85728d76bba681eb
+size 3369573916

last-checkpoint/global_step12000/zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec4ba1b6b507f1663eceb60704f806150b132d587b9f55f27741988fe2343818
+size 3149677

last-checkpoint/latest

@@ -1 +1 @@
-global_step9000
+global_step12000

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3b779b0da4fa0b940acf88ab3a1e9b5f202f72e1d1e97fdfbb99a663c3fb9108
+oid sha256:b66587b86ac7f9e8d21b13f996ded06cec0bec682dfab0f9a08702f0ae95d77d
 size 15024

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:024cda51c91a9039edefd5929d2c8e66801c9648ae3aa51499bd16e1f7543920
+oid sha256:3b12d242597b3208cff206763f66b2658a2561cee15ca32ff09d0372540fbfea
 size 15024

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c9b122e7577508b0ee5592d4271db6199d62c0e9f26bdaeb985355337f48c54
+oid sha256:7126be91d9e5b6be95962488cd837e0edc2e3b18b455506409544cb2138ce88c
 size 15024

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:259b4109c3ca03b2f8daffab8605f68685a7aaa3fde663546d748396e89c5d4b
+oid sha256:56cd20318a57445088722ed499c3d365b4f9ade8196ac71b8fc7c3278bf55f29
 size 15024

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.5183737294761532,
+  "epoch": 4.691164972634871,
   "eval_steps": 500,
-  "global_step": 9000,
+  "global_step": 12000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -25,13 +25,19 @@
       "learning_rate": 0.0001,
       "loss": 0.0949,
       "step": 9000
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 0.0001,
+      "loss": 0.0866,
+      "step": 12000
     }
   ],
   "logging_steps": 3000,
   "max_steps": 12790,
   "num_train_epochs": 5,
   "save_steps": 3000,
-  "total_flos": 118621384736768.0,
+  "total_flos": 158256349577216.0,
   "trial_name": null,
   "trial_params": null
 }