Update README.md

README.md

@@ -4,4 +4,62 @@ license: apache-2.0
 
 Slightly modified mpt-30b, which has some updates to allow gradient checkpointing/etc., to be compatible with qlora training code.
 
-Original model: https://huggingface.co/mosaicml/mpt-30b
+Original model: https://huggingface.co/mosaicml/mpt-30b
+
+My fork of qlora with mpt-30b support: https://github.com/jondurbin/qlora
+
+Differences in the qlora scripts:
+
+- requires adding `--mpt True` for mpt-based models
+- uses `--num_train_epochs` instead of `--max_steps`
+- uses airoboros prompt format (mostly 1:1 with vicuna) rather than alpaca, and expects an input file in JSONL format with "instruction" and "response"
+
+Full example of tuning (used for airoboros-mpt-30b-gpt4-1.4):
+
+```
+source /workspace/venv/bin/activate
+
+export WANDB_API_KEY=[redacted]
+export WANDB_PROJECT=airoboros-mpt-30b-gpt4-1.4
+
+python qlora.py \
+    --model_name_or_path ./mpt-30b \
+    --output_dir ./$WANDB_PROJECT-checkpoints \
+    --num_train_epochs 3 \
+    --logging_steps 1 \
+    --save_strategy steps \
+    --data_seed 11422 \
+    --save_steps 75 \
+    --save_total_limit 3 \
+    --evaluation_strategy "no" \
+    --eval_dataset_size 2 \
+    --max_new_tokens 8192 \
+    --dataloader_num_workers 3 \
+    --logging_strategy steps \
+    --remove_unused_columns False \
+    --do_train \
+    --lora_r 64 \
+    --lora_alpha 16 \
+    --lora_modules all \
+    --double_quant \
+    --quant_type nf4 \
+    --bf16 \
+    --bits 4 \
+    --warmup_ratio 0.03 \
+    --lr_scheduler_type constant \
+    --dataset ./instructions.jsonl \
+    --dataset_format airoboros \
+    --model_max_len 8192 \
+    --gradient_checkpointing \
+    --per_device_train_batch_size 6 \
+    --gradient_accumulation_steps 16 \
+    --learning_rate 0.0001 \
+    --adam_beta2 0.999 \
+    --max_grad_norm 0.3 \
+    --lora_dropout 0.05 \
+    --weight_decay 0.0 \
+    --seed 11422 \
+    --trust_remote_code \
+    --mpt True \
+    --report_to wandb
+```