Update README.md

README.md

@@ -7,11 +7,11 @@ This repo contains a low-rank adapter for LLaMA-13b fit on the Cleaned Alpaca da
 
 This version of the weights was trained with the following hyperparameters:
 
-    Cleaned dataset: Snapshot April 2, 2023
-    Epochs: 3
-    Validation set size: 2000
+    Cleaned dataset: Snapshot April 9, 2023
+    Epochs: 4
+    Validation set size: 1500
     Batch size: 128
-    Micro batch size: 8
+    Micro batch size: 4
     Cutoff length: 512
     Learning rate: 3e-4
     Lora r: 16
@@ -22,10 +22,10 @@ That is:
 python finetune.py \
     --base_model='decapoda-research/llama-13b-hf' \
     --data_path 'yahma/alpaca-cleaned' \
-    --num_epochs=3 \
+    --num_epochs=4 \
     --cutoff_len=512 \
     --output_dir='./lora-alpaca' \
     --lora_target_modules='[q_proj,k_proj, v_proj, o_proj]' \
     --lora_r=16 \
-    --val_set_size 2000 \
-    --micro_batch_size=8
+    --val_set_size 1500 \
+    --micro_batch_size=4