File size: 2,548 Bytes
1da30d6 |
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 |
#!/usr/bin/env python3
"""Download 300GB from HuggingFace datasets - FineWeb, RedPajama, etc."""
import os
import json
from datasets import load_dataset
from huggingface_hub import snapshot_download
import time
TARGET_GB = 300
OUTPUT_DIR = "/workspace/scraped_data"
os.makedirs(OUTPUT_DIR, exist_ok=True)
# Best large text datasets on HF
DATASETS = [
("HuggingFaceFW/fineweb", "sample-10BT", None), # ~50GB sample
("togethercomputer/RedPajama-Data-V2", "sample", "en_head_middle"), # ~30GB
("allenai/dolma", "v1_6-sample", None), # ~40GB
("cerebras/SlimPajama-627B", None, None), # Big
]
def get_size_gb():
total = 0
for root, dirs, files in os.walk(OUTPUT_DIR):
for f in files:
total += os.path.getsize(os.path.join(root, f))
return total / 1e9
def download_streaming(name, config, split):
"""Stream download to avoid OOM"""
print(f"\n📥 Downloading {name} ({config or 'default'})...")
try:
ds = load_dataset(name, config, split=split or "train", streaming=True, trust_remote_code=True)
shard_num = 0
batch = []
batch_size = 10000
for i, example in enumerate(ds):
text = example.get("text") or example.get("content") or str(example)
batch.append({"text": text, "source": name})
if len(batch) >= batch_size:
outfile = f"{OUTPUT_DIR}/{name.replace('/', '_')}_{shard_num:05d}.jsonl"
with open(outfile, 'w') as f:
for item in batch:
f.write(json.dumps(item) + "\n")
batch = []
shard_num += 1
size_gb = get_size_gb()
print(f" Progress: {size_gb:.1f} GB / {TARGET_GB} GB ({i:,} examples)")
if size_gb >= TARGET_GB:
print(f"✅ Target reached!")
return True
except Exception as e:
print(f" Error: {e}")
return False
if __name__ == "__main__":
print(f"🚀 Goddess HF Scraper - Target: {TARGET_GB} GB")
print(f"Output: {OUTPUT_DIR}")
start = time.time()
for name, config, split in DATASETS:
if get_size_gb() >= TARGET_GB:
break
done = download_streaming(name, config, split)
if done:
break
elapsed = time.time() - start
final_size = get_size_gb()
print(f"\n✨ Done! {final_size:.1f} GB in {elapsed/3600:.1f} hours")
|