-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathrun_real_time.sh
More file actions
26 lines (23 loc) · 989 Bytes
/
run_real_time.sh
File metadata and controls
26 lines (23 loc) · 989 Bytes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
#!/usr/bin/env bash
set -euo pipefail
TASK="press the blue button"
HLP_BASE_MODEL="/home/minji/Desktop/data/Qwen2.5-VL-7B-Instruct"
HLP_ADAPTER_PATH="/home/minji/Desktop/data/finetuned_model/ghkim/HLP_qwen_2.5_7b_QLoRA_r16_press_the_blue_button_ep60_1114_final/checkpoint-2000"
LLP_BASE_MODEL="/home/minji/Desktop/data/finetuned_model/ghkim/pi0_press_the_blue_button_ep60/030000/pretrained_model"
LLP_DATASET="/home/minji/Desktop/data/data_config/ep60_press_the_blue_button"
# LLP 관련 (필요시 config 시스템에 맞춰 수정)
MAX_STEPS=200
# ============================
python -m evaluate.eval_real_time_main \
--use_hlp=true \
--task="${TASK}" \
--max_steps="${MAX_STEPS}" \
--llp_model_path="${LLP_BASE_MODEL}" \
--dataset_repo_id="${LLP_DATASET}" \
--dataset_root="${LLP_DATASET}" \
--use_devices=true \
--hlp_model_path="${HLP_BASE_MODEL}" \
--hlp_adapter_path="${HLP_ADAPTER_PATH}" \
--hlp_use_qlora=true \
--llp_device="cuda" \
--llp_device="cuda"