Jul 20

1. 16:04

   Fix: use_cache=True, greedy decoding, model.eval() — was producing 0-char output

   AskAtul
   Verified signature

   def6c6b
2. 15:52

   Switch base image to python:3.11-slim for fresh Docker image build

   AskAtul
   Verified signature

   b97852a
3. 15:50

   Force new commit SHA with clean FP16 app.py

   AskAtul
   Verified signature

   e6c3252
4. 15:48

   Upload Dockerfile with huggingface_hub

   AskAtul
   Verified signature

   358d525
5. 15:46

   Remove load_in_8bit flag to use pure FP16 model loading

   AskAtul
   Verified signature

   ece6023
6. 15:42

   Upload requirements.txt with huggingface_hub

   AskAtul
   Verified signature

   0df7430
7. 15:40

   Use pure FP16 precision (~5.8 GB VRAM) for clean loading on Nvidia T4 GPU

   AskAtul
   Verified signature

   0a560fc
8. 15:37

   Upload requirements.txt with huggingface_hub

   AskAtul
   Verified signature

   9226cd7
9. 15:34

   Upload app.py with huggingface_hub

   AskAtul
   Verified signature

   b46992c
10. 15:34

    Upload requirements.txt with huggingface_hub

    AskAtul
    Verified signature

    903897b
11. 15:29

    Use native float16 precision for Nvidia T4 GPU (Turing architecture)

    AskAtul
    Verified signature

    f7159a3
12. 15:23

    Use standard instruction prompt and robust extraction to prevent 0-byte outputs

    AskAtul
    Verified signature

    d66da5e
13. 15:20

    Fix prompt truncation to max 1200 tokens and add decoding fallback

    AskAtul
    Verified signature

    a773699
14. 15:14

    Set pad_token_id and return clean JSON error messages

    AskAtul
    Verified signature

    7247f5e
15. 15:11

    Add detailed error reporting inside summarize endpoint

    AskAtul
    Verified signature

    28b6623
16. 15:01

    Upload app.py with huggingface_hub

    AskAtul
    Verified signature

    476b65d
17. 15:01

    Upload Dockerfile with huggingface_hub

    AskAtul
    Verified signature

    d1bdeed
18. 15:01

    Upload requirements.txt with huggingface_hub

    AskAtul
    Verified signature

    29c859c
19. 15:01

    Upload README.md with huggingface_hub

    AskAtul
    Verified signature

    1346089
20. 14:53

    Pin transformers==4.45.2 and tokenizers==0.20.0 for PyPreTokenizerTypeWrapper compatibility

    AskAtul
    Verified signature

    6b5c11b
21. 14:52

    Use official BharatGen snippet: tokenizer trust_remote_code=False & torch_dtype=bfloat16

    AskAtul
    Verified signature

    c0f7e37
22. 14:51

    Use PreTrainedTokenizerFast(tokenizer_file=tokenizer.json) directly

    AskAtul
    Verified signature

    ad33bb4
23. 14:51

    Use LlamaTokenizer directly to force pure Python tokenizer loading

    AskAtul
    Verified signature

    b5701c3
24. 14:49

    Set use_fast=False for AutoTokenizer in app.py

    AskAtul
    Verified signature

    a011140
25. 14:44

    Upload clean FastAPI app.py for transformers 4.38.2

    AskAtul
    Verified signature

    50c263f
26. 14:44

    Pin exact transformers==4.38.2 and tokenizers==0.15.2 wheel pair

    AskAtul
    Verified signature

    26f4c31
27. 14:35

    Upload README.md with huggingface_hub

    AskAtul
    Verified signature

    94bb0bc
28. 14:35

    Upload requirements.txt with huggingface_hub

    AskAtul
    Verified signature

    2d9f890
29. 14:35

    Upload app.py with huggingface_hub

    AskAtul
    Verified signature

    98ce1f9
30. 14:34

    Pin transformers==4.42.4 and accelerate==0.31.0

    AskAtul
    Verified signature

    a4ee853
31. 14:33

    Set use_safetensors=True for sharded model weights loading

    AskAtul
    Verified signature

    ac24bf5
32. 14:31

    Add safetensors>=0.4.3 to requirements.txt

    AskAtul
    Verified signature

    474eec6
33. 14:29

    Pin transformers==4.40.2 matching model config.json transformers_version

    AskAtul
    Verified signature

    0e0e8fb
34. 14:26

    Use use_fast=False for AutoTokenizer

    AskAtul
    Verified signature

    090b1fc
35. 14:26

    Pin transformers==4.35.2 with sentencepiece & fast-api

    AskAtul
    Verified signature

    ed66126
36. 14:22

    Convert app.py to FastAPI high-performance REST service

    AskAtul
    Verified signature

    856da87
37. 14:22

    Add fastapi, uvicorn, pydantic to requirements.txt

    AskAtul
    Verified signature

    3f66d00
38. 14:17

    Upload Gradio 3.50.2 app.py

    AskAtul
    Verified signature

    31dd3d6
39. 14:17

    Set sdk_version to 3.50.2 to fix Gradio 4 localhost check & jinja2 template dict error

    AskAtul
    Verified signature

    951fd3e
40. 14:14

    Pin transformers==4.40.2 to fix PyPreTokenizerTypeWrapper enum parsing in tokenizer.json

    AskAtul
    Verified signature

    83b342d
41. 14:12

    Use gr.Blocks with explicit api_name=summarize_legal_case and share=True

    AskAtul
    Verified signature

    14dbd72
42. 14:06

    Use default demo.launch() for HuggingFace Space environment

    AskAtul
    Verified signature

    25fa9de
43. 14:02

    Add server_name=0.0.0.0 and server_port=7860 to demo.launch for HF Space docker container

    AskAtul
    Verified signature

    610285d
44. 13:59

    Add api_name=summarize_legal_case to gr.Interface

    AskAtul
    Verified signature

    dae3dff
45. 13:56

    Upload README.md with huggingface_hub

    AskAtul
    Verified signature

    83f3e6e
46. 13:54

    Pin transformers==4.35.2 for compatibility with ParamBharatGen custom model code

    AskAtul
    Verified signature

    1b6c002
47. 13:52

    Set use_cache=False to completely bypass DynamicCache indexing in custom model code

    AskAtul
    Verified signature

    53c9f15
48. 13:50

    Add return_legacy_cache=True to fix DynamicCache tuple indexing in custom model code

    AskAtul
    Verified signature

    29a91ff
49. 13:48

    Upload clean scratch/app.py without shell string escaping

    AskAtul
    Verified signature

    8e3f258
50. 13:44

    Fix syntax error in app.py

    AskAtul
    Verified signature

    5779bb5