unsloth/studio/backend
Daniel Han 93ec05ced2 studio: default to UD-Q4_K_XL for GGUFs, fall back to smallest
Two changes for GGUF variant selection:

1. Default variant preference now starts with UD-Q4_K_XL (Unsloth
   Dynamic quantization) which provides better quality per bit than
   standard Q4_K_M. Also added UD-Q2_K_XL, UD-IQ2_M, UD-IQ1_M,
   UD-IQ1_S as small fallback options.

2. If the selected variant doesn't fit on disk, automatically fall
   back to the smallest GGUF variant in the repo that does fit.
   Queries all GGUF file sizes via get_paths_info() and picks the
   smallest one under the free disk space limit. If nothing fits,
   raises a clear error.

This means users with limited disk space won't get a download
error -- they'll get a smaller quantization instead.
2026-03-15 05:24:06 -07:00
..
assets Final cleanup 2026-03-12 18:28:04 +00:00
auth fix: remove old comments (#4292) 2026-03-14 16:50:13 +04:00
core studio: default to UD-Q4_K_XL for GGUFs, fall back to smallest 2026-03-15 05:24:06 -07:00
loggers Final cleanup 2026-03-12 18:28:04 +00:00
models studio: don't cap max_tokens for GGUF inference 2026-03-15 05:24:06 -07:00
plugins Final cleanup 2026-03-12 18:28:04 +00:00
requirements fix: remove old comments (#4292) 2026-03-14 16:50:13 +04:00
routes studio: don't cap max_tokens for GGUF inference 2026-03-15 05:24:06 -07:00
state Final cleanup 2026-03-12 18:28:04 +00:00
tests fix(seed): disable remote code execution in seed inspect dataset loads (#4275) 2026-03-13 19:37:43 +04:00
utils studio: default to UD-Q4_K_XL for GGUFs, fall back to smallest 2026-03-15 05:24:06 -07:00
__init__.py Final cleanup 2026-03-12 18:28:04 +00:00
colab.py Final cleanup 2026-03-12 18:28:04 +00:00
main.py fix: remove old comments (#4292) 2026-03-14 16:50:13 +04:00
run.py Final cleanup 2026-03-12 18:28:04 +00:00