apunkt/doc-to-lora
1
0
Fork 0
mirror of https://github.com/SakanaAI/doc-to-lora.git synced 2026-07-23 17:01:04 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

refactor after distillation (#6)

Browse source
This commit is contained in:
Rujikorn Charakorn 2025-07-30 18:38:51 +09:00 • committed by GitHub
parent 2b824ed813
commit 610750e6fb
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
20 changed files with 694 additions and 1648 deletions
Download patch file Download diff file Expand all files Collapse all files

8
scripts/short_ctx/gemma_qa_short_ctx_old_conf.sh
View file

@ -6,14 +6,14 @@
#SBATCH --output=slurm_logs/%x-%j.out
#SBATCH --error=slurm_logs/%x-%j.out
port=$((10000 + ($SLURM_JOBID % 50000)))
port=$((29560 + ($SLURM_JOBID % 10)))
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 \
--num_train_epochs=3 \
--per_device_train_batch_size=-1 \
--gradient_accumulation_steps=8 \
--per_device_eval_batch_size=64 \
@ -29,13 +29,11 @@ uv run accelerate launch --main_process_port $port \
--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=8192 \
--max_packed_ctx_len=16384 \
--max_packed_ctx_len=8192 \
--per_rank_gen=True \
--per_layer_processing=True \
--gen_lora_l1_reg_coef=0.1 \

6
scripts/short_ctx/gemma_qa_short_ctx_old_conf_distill.sh
View file

@ -6,14 +6,14 @@
#SBATCH --output=slurm_logs/%x-%j.out
#SBATCH --error=slurm_logs/%x-%j.out
port=$((10000 + ($SLURM_JOBID % 50000)))
port=$((29560 + ($SLURM_JOBID % 10)))
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 \
--num_train_epochs=3 \
--per_device_train_batch_size=-1 \
--gradient_accumulation_steps=16 \
--per_device_eval_batch_size=64 \
@ -29,8 +29,6 @@ uv run accelerate launch --main_process_port $port \
--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 \

42
scripts/short_ctx/gemma_qa_short_ctx_old_conf_l2l_distill.sh Normal file
View file

@ -0,0 +1,42 @@
#!/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=$((29560 + ($SLURM_JOBID % 10)))
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=3 \
--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 \
--ctx_encoder_type=per_layer_activations \
--n_latent_queries=8 \
--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 \
--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