Training in progress, step 50, checkpoint

checkpoint-50/README.md

@@ -216,4 +216,23 @@ The following `bitsandbytes` quantization config was used during training:
 ### Framework versions
 
 
+- PEFT 0.6.0
+## 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

checkpoint-50/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:45c308f44fc41d822939d4f187556b541a412450d5994c6f08b87b7670392e27
+oid sha256:81ba8e9bb46cd7431d0605620d425df31e438dcfe040d9fb7240ab631dab76d2
 size 19690328

checkpoint-50/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a77ad2b746e34edcd85ded145203c7f9f5bfc6020d4428c8bb32b4342cb76c34
+oid sha256:e883ceaf3c69791c4e13ae3e58e571eaee58054ab5a3846b611c5b2b7887b9e4
 size 38087098

checkpoint-50/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8b81ff2e6c04e9ac576799d077071b8138451e3db191535bd96291dd58f15897
+oid sha256:5f222629472477b761a77581c63b2d54d08cc9afcfee220260faf8cc8bde627d
 size 14244

checkpoint-50/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:73c3573e9aef7d06d0822177200fa67c2fd9c07144cebfce2de7c64fa8186d55
+oid sha256:69f0be3d3f244f6ed78a99ef73024bb1cc874b8d8c497ea6ae16ba6bbb24fdde
 size 1064

checkpoint-50/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.4239814877510071,
+  "best_metric": 0.4395545721054077,
   "best_model_checkpoint": "output/checkpoint-50",
-  "epoch": 0.7220216606498195,
+  "epoch": 0.1353637901861252,
   "eval_steps": 50,
   "global_step": 50,
   "is_hyper_param_search": false,
@@ -9,32 +9,32 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.01,
+      "epoch": 0.0,
       "learning_rate": 0.0,
-      "loss": 0.3991,
+      "loss": 0.9268,
       "step": 1
     },
     {
-      "epoch": 0.72,
-      "learning_rate": 1.8902439024390243e-05,
-      "loss": 0.3172,
+      "epoch": 0.14,
+      "learning_rate": 6.111111111111112e-06,
+      "loss": 0.5897,
       "step": 50
     },
     {
-      "epoch": 0.72,
-      "eval_accuracy": 0.84,
-      "eval_loss": 0.4239814877510071,
-      "eval_runtime": 28.6347,
-      "eval_samples_per_second": 6.111,
-      "eval_steps_per_second": 1.537,
+      "epoch": 0.14,
+      "eval_accuracy": 0.8628571428571429,
+      "eval_loss": 0.4395545721054077,
+      "eval_runtime": 26.2003,
+      "eval_samples_per_second": 6.679,
+      "eval_steps_per_second": 1.679,
       "step": 50
     }
   ],
   "logging_steps": 50,
-  "max_steps": 200,
+  "max_steps": 800,
   "num_train_epochs": 3,
   "save_steps": 50,
-  "total_flos": 1.908269520728064e+16,
+  "total_flos": 1.83883393431552e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-50/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6e1226be91d4f2bcdcde81b49645b5882b397e67c3f85a88d2c0a49245bf8c5
+oid sha256:eab90039203ebde7fc5537527df9e43bcc8971062162dee256a3fd06a20fcd44
 size 4600