File size: 4,779 Bytes
df529cc | 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 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 | #!/usr/bin/env python3
"""Generate a few ChartQA responses and audit the DyME CoT format."""
from __future__ import annotations
import argparse
import json
import sys
from pathlib import Path
import torch
from PIL import Image
from transformers import AutoProcessor, LlavaOnevisionForConditionalGeneration
_PROJECT_ROOT = Path(__file__).resolve().parent.parent
if str(_PROJECT_ROOT) not in sys.path:
sys.path.insert(0, str(_PROJECT_ROOT))
from data_utils.chart.data_collector import prepare_chart_rl_data
SECTIONS = ("Goal:", "Observation:", "Reasoning:", "Conclusion:", "Answer:")
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser()
parser.add_argument("--model_path", required=True)
parser.add_argument(
"--processor_path",
default=None,
help="Tokenizer/processor source; defaults to --model_path.",
)
parser.add_argument("--dataset", required=True)
parser.add_argument("--indices", default="0,1,100,1000")
parser.add_argument("--max_new_tokens", type=int, default=300)
parser.add_argument("--output", default=None)
return parser.parse_args()
def ordered_sections(text: str) -> bool:
positions = [text.find(section) for section in SECTIONS]
return all(position >= 0 for position in positions) and positions == sorted(positions)
def main() -> int:
args = parse_args()
indices = [int(value.strip()) for value in args.indices.split(",") if value.strip()]
rows = prepare_chart_rl_data(args.dataset)
if not indices or min(indices) < 0 or max(indices) >= len(rows):
raise ValueError(f"indices must be within [0, {len(rows) - 1}]")
model_path = str(Path(args.model_path).resolve())
processor_path = str(Path(args.processor_path or args.model_path).resolve())
processor = AutoProcessor.from_pretrained(processor_path, local_files_only=True)
processor.tokenizer.padding_side = "left"
model = LlavaOnevisionForConditionalGeneration.from_pretrained(
model_path,
torch_dtype=torch.bfloat16,
low_cpu_mem_usage=True,
attn_implementation="sdpa",
local_files_only=True,
).to("cuda:0")
model.eval()
results = []
for index in indices:
row = rows[index]
image = Image.open(row["image"]).convert("RGB")
messages = [
{
"role": "user",
"content": [
{"type": "image"},
{"type": "text", "text": row["prompt"]},
],
}
]
text = processor.apply_chat_template(
messages, add_generation_prompt=True, tokenize=False
)
inputs = processor(text=[text], images=[image], return_tensors="pt")
inputs = {key: value.to("cuda:0") for key, value in inputs.items()}
prompt_length = inputs["input_ids"].shape[1]
with torch.inference_mode():
output_ids = model.generate(
**inputs,
max_new_tokens=args.max_new_tokens,
do_sample=False,
repetition_penalty=1.0,
use_cache=True,
)
generated_ids = output_ids[:, prompt_length:]
response = processor.batch_decode(
generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False
)[0].strip()
result = {
"index": index,
"question": row["question_wo_prompt"],
"gold": row["reference_answer"],
"response": response,
"ordered_cot_format": ordered_sections(response),
"response_words": len(response.split()),
}
results.append(result)
print(
f"\n[COT sample {index}] format={result['ordered_cot_format']} "
f"words={result['response_words']} gold={result['gold']}\n{response}",
flush=True,
)
format_count = sum(result["ordered_cot_format"] for result in results)
summary = {
"model_path": model_path,
"processor_path": processor_path,
"samples": len(results),
"ordered_cot_count": format_count,
"ordered_cot_rate": format_count / len(results),
"mean_response_words": sum(item["response_words"] for item in results) / len(results),
"results": results,
}
print("\n[COT summary] " + json.dumps({
key: value for key, value in summary.items() if key != "results"
}, ensure_ascii=False), flush=True)
if args.output:
output_path = Path(args.output)
output_path.parent.mkdir(parents=True, exist_ok=True)
output_path.write_text(json.dumps(summary, ensure_ascii=False, indent=2), encoding="utf-8")
return 0
if __name__ == "__main__":
raise SystemExit(main())
|