Training in progress, step 150, checkpoint

checkpoint-150/README.md

@@ -216,4 +216,23 @@ The following `bitsandbytes` quantization config was used during training:
 ### Framework versions
 
 
+- PEFT 0.6.0.dev0
+## Training procedure
+
+
+The following `bitsandbytes` quantization config was used during training:
+- quant_method: bitsandbytes
+- load_in_8bit: True
+- load_in_4bit: False
+- llm_int8_threshold: 6.0
+- llm_int8_skip_modules: None
+- llm_int8_enable_fp32_cpu_offload: False
+- llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: fp4
+- bnb_4bit_use_double_quant: False
+- bnb_4bit_compute_dtype: float32
+
+### Framework versions
+
+
 - PEFT 0.6.0.dev0

checkpoint-150/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a68637fc3c68aba0871f07d2a1e14f18a03f24a602f068d993ed99ab9a92ff3b
+oid sha256:0edb26a785cc8cf4e931d5d46fe98aefc38569f336d6b308468931bf782ab245
 size 9873829

checkpoint-150/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c14c1e50a0ceb905ffc78467e51a78a4ac123a50774cd7607150c94788671c3b
+oid sha256:255c211b958a2fc9d60e541f256ab81db30cc762cbc041190980e97472f0f870
 size 42724

checkpoint-150/trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.6410807371139526,
+  "best_metric": 0.6430063247680664,
   "best_model_checkpoint": "output/checkpoint-50",
   "epoch": 0.12025253031365868,
   "eval_steps": 50,
@@ -11,52 +11,52 @@
     {
       "epoch": 0.0,
       "learning_rate": 4.444444444444445e-07,
-      "loss": 0.5558,
+      "loss": 0.5542,
       "step": 1
     },
     {
       "epoch": 0.04,
       "learning_rate": 2e-05,
-      "loss": 0.5679,
+      "loss": 0.554,
       "step": 50
     },
     {
       "epoch": 0.04,
-      "eval_accuracy": 0.6776119402985075,
-      "eval_loss": 0.6410807371139526,
-      "eval_runtime": 61.731,
-      "eval_samples_per_second": 5.427,
-      "eval_steps_per_second": 1.361,
+      "eval_accuracy": 0.6626865671641791,
+      "eval_loss": 0.6430063247680664,
+      "eval_runtime": 61.7692,
+      "eval_samples_per_second": 5.423,
+      "eval_steps_per_second": 1.36,
       "step": 50
     },
     {
       "epoch": 0.08,
       "learning_rate": 4.222222222222222e-05,
-      "loss": 0.6108,
+      "loss": 0.5957,
       "step": 100
     },
     {
       "epoch": 0.08,
-      "eval_accuracy": 0.6925373134328359,
-      "eval_loss": 0.6733226776123047,
-      "eval_runtime": 56.5238,
-      "eval_samples_per_second": 5.927,
-      "eval_steps_per_second": 1.486,
+      "eval_accuracy": 0.6686567164179105,
+      "eval_loss": 0.6705919504165649,
+      "eval_runtime": 60.4388,
+      "eval_samples_per_second": 5.543,
+      "eval_steps_per_second": 1.39,
       "step": 100
     },
     {
       "epoch": 0.12,
       "learning_rate": 6.444444444444446e-05,
-      "loss": 0.5865,
+      "loss": 0.5668,
       "step": 150
     },
     {
       "epoch": 0.12,
-      "eval_accuracy": 0.6865671641791045,
-      "eval_loss": 0.7122625708580017,
-      "eval_runtime": 61.5948,
-      "eval_samples_per_second": 5.439,
-      "eval_steps_per_second": 1.364,
+      "eval_accuracy": 0.6805970149253732,
+      "eval_loss": 0.7224913835525513,
+      "eval_runtime": 56.4882,
+      "eval_samples_per_second": 5.93,
+      "eval_steps_per_second": 1.487,
       "step": 150
     }
   ],

checkpoint-150/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e2cf806244b71c753bc4b76d2695973243d02590ea5a24f377343d4ee219e984
+oid sha256:87bd30bd4b21b4a0c66f1fd545858ef6e2d33b5984a8d292e79011acc3affd6b
 size 4472