Model fine-tuning

[carlosgjs]

• Fine-tune: Training/validation data loading #16
• Fine-tune: Training script using LORA #17
• Fine-tune: Experimentation across parameters (quant mode, LORA: dropout, rank, alpha, etc.) #18
• Fine-tune: Experimentation across base models (llama-7B, llama-13B, codellama-7B, codellama-13B) #19
• Fine-tune: Model runtime optimization #20