Model save

README.md

@@ -41,10 +41,10 @@ The following hyperparameters were used during training:
 - eval_batch_size: 8
 - seed: 42
 - distributed_type: multi-GPU
-- num_devices: 2
+- num_devices: 4
 - gradient_accumulation_steps: 8
-- total_train_batch_size: 16
-- total_eval_batch_size: 16
+- total_train_batch_size: 32
+- total_eval_batch_size: 32
 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_steps: 100
@@ -58,6 +58,6 @@ The following hyperparameters were used during training:
 
 - PEFT 0.16.0
 - Transformers 4.53.2
-- Pytorch 2.5.1+cu124
+- Pytorch 2.5.1
 - Datasets 4.0.0
 - Tokenizers 0.21.2

final/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3f0b47a4fa6dc132dd687e758ebb68bea4e8c6bdc72c7b5136aadedeb098546c
+oid sha256:ad27aa3603b1e3d6564ff8dc7a9ebcbfe18cbc8063d4ae74ca4a342ec38eddab
 size 4373064

final/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d2bbbd2f91a551fa0e22a065002a0d9b12f4155c7518ac28f5715d6868944594
+oid sha256:a0d3008385bee87dba0c5283ff6fd6bbbc4a7a71655a07a9c4dee3da83b10b39
 size 6520