style: apply ruff formatting to structured_summarizer
bd7d2c1
ming
Claude
commited on
Spaces:
Running
Running
Commit History
docs: add comprehensive V4 API documentation and optimize inference with SDPA
0072188
ming
Claude
commited on
chore: code formatting improvements and update gitignore
db3b809
ming
commited on
Add V4 local server setup with MPS optimization for Android testing
45b6536
ming
Claude
commited on
Apply ruff formatting to Loguru migration files
75fe59b
ming
Claude
commited on
Migrate logging from stdlib to Loguru for structured logging
7ab470d
ming
Claude
commited on
Remove outlines library and all related code
d25a17f
ming
commited on
Migrate to Ruff for linting/formatting and add comprehensive import tests
29ed661
ming
commited on
Fix Outlines API usage for V4 JSON streaming endpoint
b47201f
ming
commited on
Fix Outlines API usage - handle different calling patterns
4c4036e
ming
commited on
Fix Outlines import - use generator instead of generate
9452571
ming
commited on
Change Outlines debug log to info level for visibility
5fa0ba2
ming
commited on
Add detailed Outlines API exploration and logging
f099d0c
ming
commited on
Fix Outlines import - use correct API for installed version
e8ab865
ming
commited on
Improve Outlines import error handling and logging
33cd483
ming
commited on
Improve error messaging for Outlines unavailability
734e281
ming
commited on
Fix Python 3.10 requirement and torch_dtype deprecation
6b859f2
ming
commited on
Add Outlines JSON streaming endpoint for V4 structured summarization
441f66b
ming
commited on
Fix JSON parsing errors in V4 NDJSON stream
85dcd04
ming
commited on
fix: Move inputs to model device in _single_chunk_summarize to fix CPU/GPU device mismatch
cfe8d29
ming
commited on
Fix device placement error in V2 BART model
12a2e7c
ming
commited on
Implement Option 3: Use FP16 for 2-3x faster inference
7fff563
ming
commited on
Optimize V4 generation speed: greedy decoding + reduced max_tokens
fd2a8c1
ming
commited on
Fix buffer parsing and strengthen brevity constraints
d112a13
ming
commited on
Optimize V4 output verbosity and generation speed
17499f7
ming
commited on
Fix bitsandbytes UID error with getpass patch
dd29a6d
ming
commited on
Switch V4 to GPU INT4 quantization with Qwen-1.5B
a36f560
ming
commited on
debug: Add comprehensive logging to diagnose 4-token issue
df75294
ming
commited on
feat: Guarantee complete V4 NDJSON summaries with fallback
b321440
ming
commited on
fix: Use Qwen chat template and harden NDJSON parsing
bf21a65
ming
commited on
feat: Switch V4 to Qwen2.5-1.5B for HF memory compatibility
fe47248
ming
commited on
perf: Disable V2 warmup to save memory for V4
1b76b21
ming
commited on
feat: Change device_map to auto for V4 model
d0701b0
ming
commited on
feat: Switch V4 model to Phi-3-mini for better structured output
7019b66
ming
commited on
feat: Add V4 NDJSON patch-based structured summarization
93c9664
ming
commited on
Revert adaptive token logic, restore client-controlled max_tokens
6a1e8a3
ming
Claude
commited on
fix: Backend ignores client max_tokens to verify Android app hypothesis
80ea70f
ming
Claude
commited on
fix: CRITICAL - Override model config defaults causing early stopping
6c96c54
ming
Claude
commited on
fix: Improve V3 summary completeness with enhanced token allocation
6b2de93
ming
Claude
commited on
fix: V3 API mid-sentence cutoff with adaptive token calculation
5e83010
ming
Claude
commited on
Fix V3 API to support both URL and text input
f724bab
ming
Claude
commited on
feat: Implement V3 Web Scraping + Summarization API
2ed2bd7
ming
commited on
Update v2 API and HF streaming summarizer improvements
fc9914e
ming
commited on
Improve V2 summarization: adaptive tokens, recursive summarization, better defaults
9884884
ming
commited on
Comment out asyncio.sleep for better streaming performance
698636a
ming
commited on
Enhance HF V2 summaries: increase token limits, improve length parameters
52f6c42
ming
commited on
Improve V2 summarization prompts for more detailed output
de8b12b
ming
commited on
Add model ID logging to HF streaming service
e186c67
ming
commited on
Fix HF streaming beam expansion issue
df1ab37
ming
commited on
Enforce singleton batch for HF streaming (fix TextStreamer batch size 1 error)
843f837
ming
commited on