Update README.md

README.md

@@ -25,13 +25,13 @@ accelerate launch $BASE_DIR/qlora/train.py \
   --num_train_epochs 1 \
   --logging_steps 1 \
   --save_strategy steps \
-  --save_steps 120 \
+  --save_steps 75 \
   --save_total_limit 2 \
   --data_seed 11422 \
   --evaluation_strategy steps \
   --per_device_eval_batch_size 4 \
   --eval_dataset_size 0.01 \
-  --eval_steps 120 \
+  --eval_steps 75 \
   --max_new_tokens 1024 \
   --dataloader_num_workers 3 \
   --logging_strategy steps \
@@ -47,16 +47,15 @@ accelerate launch $BASE_DIR/qlora/train.py \
   --dataset habanoz/airoboros-3.1-no-mathjson-max-1k \
   --dataset_format airoboros_chat \
   --model_max_len 1024 \
-  --per_device_train_batch_size 1 \
-  --gradient_accumulation_steps 16 \
+  --per_device_train_batch_size 4 \
+  --gradient_accumulation_steps 4 \
   --learning_rate 1e-5 \
   --adam_beta2 0.999 \
   --max_grad_norm 0.3 \
   --lora_dropout 0.0 \
   --weight_decay 0.0 \
   --seed 11422 \
-  --gradient_checkpointing False \
+  --gradient_checkpointing \
   --use_flash_attention_2 \
-  --ddp_find_unused_parameters False \
-  --trust_remote_code True
+  --ddp_find_unused_parameters False
   ```