Add/update README

README.md

@@ -0,0 +1,65 @@
+# medmcqa LoRA Models
+
+This repository contains LoRA (Low-Rank Adaptation) models trained on the medmcqa dataset.
+
+## Models in this repository:
+
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0002_data_size1000_max_steps=500_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0002_data_size1000_max_steps=500_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0001_data_size1000_max_steps=100_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0001_data_size1000_max_steps=100_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr5e-06_data_size1000_max_steps=500_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr5e-06_data_size1000_max_steps=500_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0003_data_size1000_max_steps=500_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0003_data_size1000_max_steps=500_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr1e-05_data_size1000_max_steps=500_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr1e-05_data_size1000_max_steps=500_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0001_data_size1000_max_steps=500_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0001_data_size1000_max_steps=500_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0002_data_size1000_max_steps=100_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0002_data_size1000_max_steps=100_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0003_data_size1000_max_steps=100_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr0.0003_data_size1000_max_steps=100_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr1e-05_data_size1000_max_steps=1000_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr1e-05_data_size1000_max_steps=1000_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr5e-06_data_size1000_max_steps=1000_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr5e-06_data_size1000_max_steps=1000_seed=123
+- `llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr5e-05_data_size1000_max_steps=500_seed=123/`: LoRA adapter for llama_finetune_medmcqa_r16_alpha=32_dropout=0.05_lr5e-05_data_size1000_max_steps=500_seed=123
+
+## Usage
+
+To use these LoRA models, you'll need the `peft` library:
+
+```bash
+pip install peft transformers torch
+```
+
+Example usage:
+
+```python
+from peft import PeftModel
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+# Load base model
+base_model_name = "your-base-model"  # Replace with actual base model
+model = AutoModelForCausalLM.from_pretrained(base_model_name)
+tokenizer = AutoTokenizer.from_pretrained(base_model_name)
+
+# Load LoRA adapter
+model = PeftModel.from_pretrained(
+    model, 
+    "supergoose/medmcqa",
+    subfolder="model_name_here"  # Replace with specific model folder
+)
+
+# Use the model
+inputs = tokenizer("Your prompt here", return_tensors="pt")
+outputs = model.generate(**inputs)
+```
+
+## Training Details
+
+- Dataset: medmcqa
+- Training framework: LoRA/PEFT
+- Models included: 11 variants
+
+## Files Structure
+
+Each model folder contains:
+- `adapter_config.json`: LoRA configuration
+- `adapter_model.safetensors`: LoRA weights
+- `tokenizer.json`: Tokenizer configuration
+- Additional training artifacts
+
+---
+*Generated automatically by LoRA uploader script*