Training in progress, step 100, checkpoint

checkpoint-100/README.md

@@ -216,23 +216,4 @@ 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-100/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f68ef74d88fadddbf3434d91fee124280668f80a340fb0f5cfd4c60a0f462b00
+oid sha256:3f9ede17267dcaedaf243ee0c9dfaec7a9839db83a912ad3db13a8a8fd47f65c
 size 9873829

checkpoint-100/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:79b16eb3ee3d773dc55a3001899da84ed5cfc088b7b1ff81f75ed29f468b81be
+oid sha256:7a261d85a9162739496c39778cc079ee8abc4d3488d8fdc09bc4103dda68bda2
 size 42724

checkpoint-100/trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.6410807371139526,
+  "best_metric": 0.6430063247680664,
   "best_model_checkpoint": "output/checkpoint-50",
   "epoch": 0.08016835354243912,
   "eval_steps": 50,
@@ -11,37 +11,37 @@
     {
       "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
     }
   ],

checkpoint-100/training_args.bin

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