Training in progress, step 4000, checkpoint

checkpoint-4000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d642a4a413c0a0de63898789e5d3cc282dfdcf14f73ea4d57977e3936a52bb9c
+oid sha256:0d7d747a9b573bd9b07df13f065931d9361054f88adb7b6c9802f565177dcd49
 size 51007160

checkpoint-4000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:089952b84a7823f6adf250f989d3e977edb19e1260db73b02caef88cbe18a9f0
+oid sha256:ae50cb9a0c4ce02865f7c74c9ec6e69d16e879f7438cc610845a4979c17bf01d
 size 102078202

checkpoint-4000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:646b9668d3f9e0223ab52454b10f0a9f05db560c41d95e68e9d4441ff5ccbc4e
+oid sha256:d2d600a948988d9b48358f98258c801dc3497028ae53b9a467e56c11131c2d24
 size 14308

checkpoint-4000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 6.568221092224121,
-  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/en_clm/wikipedia_30/checkpoint-4000",
-  "epoch": 2.1036024191427822,
+  "best_metric": 5.8541717529296875,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/fr_clm/wikipedia_30/checkpoint-4000",
+  "epoch": 4.294149221685453,
   "eval_steps": 2000,
   "global_step": 4000,
   "is_hyper_param_search": false,
@@ -9,33 +9,33 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0518012095713911,
-      "eval_loss": 7.518920421600342,
-      "eval_runtime": 2.2399,
-      "eval_samples_per_second": 1422.808,
-      "eval_steps_per_second": 89.288,
+      "epoch": 2.1470746108427265,
+      "eval_loss": 7.138043403625488,
+      "eval_runtime": 0.8415,
+      "eval_samples_per_second": 1273.93,
+      "eval_steps_per_second": 79.621,
       "step": 2000
     },
     {
-      "epoch": 2.1036024191427822,
-      "grad_norm": 1.2947496175765991,
+      "epoch": 4.294149221685453,
+      "grad_norm": 1.455039381980896,
       "learning_rate": 1e-05,
-      "loss": 7.6141,
+      "loss": 7.213,
       "step": 4000
     },
     {
-      "epoch": 2.1036024191427822,
-      "eval_loss": 6.568221092224121,
-      "eval_runtime": 2.1236,
-      "eval_samples_per_second": 1500.783,
-      "eval_steps_per_second": 94.182,
+      "epoch": 4.294149221685453,
+      "eval_loss": 5.8541717529296875,
+      "eval_runtime": 0.7753,
+      "eval_samples_per_second": 1382.676,
+      "eval_steps_per_second": 86.417,
       "step": 4000
     }
   ],
   "logging_steps": 4000,
   "max_steps": 100000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 53,
+  "num_train_epochs": 108,
   "save_steps": 4000,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -49,7 +49,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1034022628884480.0,
+  "total_flos": 1034006468493312.0,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

checkpoint-4000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ab68434ec7156d3a63e4783f824871a48d62c7c2fdcc831fdd5471c5f4aec7fe
+oid sha256:26102d0ac750ce531ebda9d7577ccea6c8ecbc93b04bdff23226e376dd5609a2
 size 5368