Training in progress, step 150, checkpoint

checkpoint-150/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f7bce82964ec1e57f66e49a5f3a26133573e38332ce871bc571d4d9e4130cfca
+oid sha256:ce3ce313396f6b85355d4432b51fd87e624b25ea26bc55ef592748513912cc12
 size 19690328

checkpoint-150/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ea2d23b0e3e4b644ad51b9f60c43c564aab750d082ec0003a2426d886caa1c96
+oid sha256:750ad12ac1b375539b663c6b3979d2401ba3a700d12cc2e7181d6023fc123f28
 size 38087098

checkpoint-150/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d4062dece82eaba8fea3f217476a7ecbd0af4a7973f0c6ee8c617903a8b756b
+oid sha256:598108331c32c4d5d147ffec469be0ca218f58188a4b36970f4bd18efb1e0a80
 size 14244

checkpoint-150/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f178ac6ba62c0610ca3639f96006f7544b9bd0127efed4bdc19dd8aa7df0cace
+oid sha256:014f4bcfecab0d20148f3ce1dc4fbec63ea785bfa45b6765fba42b0ad38c087c
 size 1064

checkpoint-150/special_tokens_map.json

@@ -2,6 +2,6 @@
   "bos_token": "<s>",
   "eos_token": "</s>",
   "pad_token": "<pad>",
-  "sep_token": "\n\n",
+  "sep_token": "<s>",
   "unk_token": "<unk>"
 }

checkpoint-150/tokenizer_config.json

@@ -39,7 +39,7 @@
   "eos_token": "</s>",
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<pad>",
-  "sep_token": "\n\n",
+  "sep_token": "<s>",
   "tokenizer_class": "BloomTokenizer",
   "unk_token": "<unk>"
 }

checkpoint-150/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.6207379102706909,
-  "best_model_checkpoint": "output/checkpoint-100",
-  "epoch": 0.6259780907668232,
+  "best_metric": 0.47385331988334656,
+  "best_model_checkpoint": "output/checkpoint-150",
+  "epoch": 0.12316534948167915,
   "eval_steps": 50,
   "global_step": 150,
   "is_hyper_param_search": false,
@@ -11,60 +11,60 @@
     {
       "epoch": 0.0,
       "learning_rate": 0.0,
-      "loss": 0.4694,
+      "loss": 0.8903,
       "step": 1
     },
     {
-      "epoch": 0.21,
-      "learning_rate": 1.25e-05,
-      "loss": 0.3006,
+      "epoch": 0.04,
+      "learning_rate": 1.2222222222222224e-05,
+      "loss": 0.7189,
       "step": 50
     },
     {
-      "epoch": 0.21,
-      "eval_accuracy": 0.6971428571428572,
-      "eval_loss": 0.6772400140762329,
-      "eval_runtime": 28.6241,
-      "eval_samples_per_second": 6.114,
-      "eval_steps_per_second": 1.537,
+      "epoch": 0.04,
+      "eval_accuracy": 0.8228571428571428,
+      "eval_loss": 0.5415178537368774,
+      "eval_runtime": 28.5756,
+      "eval_samples_per_second": 6.124,
+      "eval_steps_per_second": 1.54,
       "step": 50
     },
     {
-      "epoch": 0.42,
-      "learning_rate": 1.8597560975609757e-05,
-      "loss": 0.1715,
+      "epoch": 0.08,
+      "learning_rate": 1.8658536585365855e-05,
+      "loss": 0.5723,
       "step": 100
     },
     {
-      "epoch": 0.42,
-      "eval_accuracy": 0.8228571428571428,
-      "eval_loss": 0.6207379102706909,
-      "eval_runtime": 28.9688,
-      "eval_samples_per_second": 6.041,
-      "eval_steps_per_second": 1.519,
+      "epoch": 0.08,
+      "eval_accuracy": 0.8,
+      "eval_loss": 0.49520450830459595,
+      "eval_runtime": 28.516,
+      "eval_samples_per_second": 6.137,
+      "eval_steps_per_second": 1.543,
       "step": 100
     },
     {
-      "epoch": 0.63,
-      "learning_rate": 1.554878048780488e-05,
-      "loss": 0.1573,
+      "epoch": 0.12,
+      "learning_rate": 1.5609756097560978e-05,
+      "loss": 0.5122,
       "step": 150
     },
     {
-      "epoch": 0.63,
-      "eval_accuracy": 0.76,
-      "eval_loss": 0.7828966379165649,
-      "eval_runtime": 26.1173,
-      "eval_samples_per_second": 6.701,
-      "eval_steps_per_second": 1.685,
+      "epoch": 0.12,
+      "eval_accuracy": 0.7885714285714286,
+      "eval_loss": 0.47385331988334656,
+      "eval_runtime": 28.2004,
+      "eval_samples_per_second": 6.206,
+      "eval_steps_per_second": 1.56,
       "step": 150
     }
   ],
   "logging_steps": 50,
   "max_steps": 400,
-  "num_train_epochs": 2,
+  "num_train_epochs": 1,
   "save_steps": 50,
-  "total_flos": 5.565524507676672e+16,
+  "total_flos": 5.561965309452288e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-150/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5027c9d6c3122f21d433b1e532b4aebee600801d0d4723deb6d4671b22b3d2da
+oid sha256:090099aa07c672025c60437897a7a19f63d8d7bd4284af2562fc632695431956
 size 4600