Add distillation training (#5)

- work with vllm self-gen data (everything is tokenized during self-gen)
- fix chat template!
- not use liger kernel for distillation (for both training and eval)

scripts/short_ctx/gemma_qa_short_ctx_old_conf_distill.sh

@@ -0,0 +1,43 @@
+#!/bin/bash
+#SBATCH --job-name=ctxlora
+#SBATCH --nodes=1
+#SBATCH --partition=a3
+#SBATCH --gpus=4
+#SBATCH --output=slurm_logs/%x-%j.out
+#SBATCH --error=slurm_logs/%x-%j.out
+
+port=$((10000 + ($SLURM_JOBID % 50000)))
+echo "Using port: $port"
+
+# --gradient_accumulation_steps=8 --gradient_clipping=1.0
+uv run accelerate launch --main_process_port $port \
+--num_processes=4 --gpu_ids all intx_sft.py $1 \
+--model_name_or_path=google/gemma-2-2b-it \
+--num_train_epochs=5 \
+--per_device_train_batch_size=-1 \
+--gradient_accumulation_steps=16 \
+--per_device_eval_batch_size=64 \
+--target_modules=down_proj \
+--num_blocks=9 \
+--num_self_attn_per_block=0 \
+--n_latent_queries=208 \
+--num_pre_head_layers=1 \
+--lora_r=8 \
+--eval_strategy=no \
+--eval_steps=1000 \
+--save_steps=1000 \
+--learning_rate=4e-5 \
+--lora_dropout=0.0 \
+--neftune_noise_alpha=5 \
+--add_negative_prompt=False \
+--add_repeat_prompt=False \
+--use_sequence_packing=True \
+--max_qas_len=2048 \
+--max_qas_per_sample=1 \
+--max_packed_inp_len=4096 \
+--max_packed_ctx_len=4096 \
+--per_rank_gen=True \
+--per_layer_processing=True \
+--gen_lora_l1_reg_coef=0.1 \
+--logging_steps=50 \
+--use_kl_loss=True