theroyallab/tabbyAPI
1
0
Fork 0
mirror of https://github.com/theroyallab/tabbyAPI.git synced 2026-05-11 00:10:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'origin/main'

Browse Source
This commit is contained in:
turboderp
2026-05-09 23:18:52 +02:00
parent 4a8cb08a24 54c1e56019
commit 5d964494b6
2 changed files with 9 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
config_sample.yml
View File

@@ -166,10 +166,10 @@ model:
# Do NOT enable this if the model is not a reasoning model (e.g. deepseek-r1 series)
reasoning: false
# The start token for reasoning conetnt (default: "<think>")
# The start token for reasoning content (default: "<think>")
reasoning_start_token: "<think>"
# The end token for reasoning conetnt (default: "</think>")
# The end token for reasoning content (default: "</think>")
reasoning_end_token: "</think>"
# Suppress this text whenever it appears in the beginning of a reasoning block (default: None)

9
docker/Dockerfile
View File

@@ -26,8 +26,13 @@ WORKDIR /app
# Get requirements
COPY pyproject.toml .
# Install packages specified in pyproject.toml cu12, extras
RUN pip install --no-cache-dir .[cu12,extras]
# Install cu12 group first — pins torch+cu128, exllamav2/v3+cu128, flash_attn+cu128.
# The 'extras' group (infinity-emb, sentence-transformers) is installed separately
# with --no-deps so pip cannot resolve xformers transitively and pull a cu130 wheel,
# which would cause libcudart.so.13 ImportError on driver 590.x (cu128-only hosts).
# See: https://github.com/theroyallab/tabbyAPI/issues/414
RUN pip install --no-cache-dir .[cu12]
RUN pip install --no-cache-dir --no-deps .[extras]
RUN rm pyproject.toml