1# Large File Uploads
2 
3The standard `client.files.create()` silently fails on JSONL files >~150MB (Azure returns 500 during job execution). Use the chunked Uploads API:
4 
5```python
6upload = client.uploads.create(filename="data.jsonl", purpose="fine-tune", bytes=file_size, mime_type="application/jsonl")
7part_ids = []
8with open(filepath, "rb") as f:
9    while chunk := f.read(64 * 1024 * 1024):  # 64MB chunks
10        part = client.uploads.parts.create(upload_id=upload.id, data=chunk)
11        part_ids.append(part.id)
12completed = client.uploads.complete(upload_id=upload.id, part_ids=part_ids)
13file_id = completed.file.id
14```
15 
16**Important:** Requires `openai.AzureOpenAI()` client, NOT `openai.OpenAI()` with `/v1/` URL. The project endpoint returns 404 for upload operations.
17 
18| File Size | Method |
19|-----------|--------|
20| < 100MB | Standard `files.create()` |
21| 100MB–5GB | Chunked Uploads API |
22| > 5GB | Split dataset |
23