ochafik
Follow
Popular repositories Loading
-
-
openscad-playground
openscad-playground PublicForked from openscad/openscad-playground
OpenSCAD Web Playground
-
optimized-loops-macros
optimized-loops-macros PublicMacro-powered loop optimizations (Range for comprehensions)
Scala 14
-
rpi-raspbian-bazel
rpi-raspbian-bazel PublicBazel Raspbian build script + off-the-shelf resin/rpi-raspbian Docker image derivative.
831 contributions in the last year
Less
No contributions.
Low contributions.
Medium-low contributions.
Medium-high contributions.
High contributions.
More
Contribution activity
March 2025
Created 9 commits in 2 repositories
Created 3 repositories
-
ochafik/cortex.cpp
C++
This contribution was made on Mar 16
-
ochafik/rl-experiments
Jupyter Notebook
This contribution was made on Mar 7
-
ochafik/gorilla
Python
This contribution was made on Mar 5
Created a pull request in ggml-org/llama.cpp that received 12 comments
server: streaming of tool calls and thoughts when --jinja is on
This PR is still WIP (see todos at the bottom) but welcoming early feedback / testing Support streaming of tool calls in OpenAI format Improve han…
+3,616
−946
lines changed
•
12
comments
Opened 9 other pull requests in 3 repositories
ggml-org/llama.cpp
7
merged
-
server: extract <think> tags from qwq outputsThis contribution was made on Mar 10 -
tool-call: allow missing content in message if tool_calls providedThis contribution was made on Mar 10 -
tool-call: ensure there's always a non-empty tool call idThis contribution was made on Mar 10 -
sampler: fixes trigger tokens + lazy grammars (fix typo cast from token to string)This contribution was made on Mar 9 -
sync: minja (support QwQ-32B)This contribution was made on Mar 6 -
docs: update function-calling.md w/ template override needed by functionary-small-v3.2
This contribution was made on Mar 6
-
server: fix response_format w/ json_schema.schemaThis contribution was made on Mar 3
Telosnex/llama.cpp
1
merged
-
Fixes for phi-4 support
This contribution was made on Mar 9
google/minja
1
merged
-
Support Qwen/QwQ-32B (needed str.{split,lstrip})
This contribution was made on Mar 6
Reviewed 6 pull requests in 1 repository
ggml-org/llama.cpp
6 pull requests
-
tool-call: Phi-4 supportThis contribution was made on Mar 16 -
server: extract <think> tags from qwq outputsThis contribution was made on Mar 10 -
sampler: fixes trigger tokens + lazy grammars (fix typo cast from token to string)This contribution was made on Mar 10 -
tool-call: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammarsThis contribution was made on Mar 5 -
main: allow preloading conversation with -p and add -st / --single-turn
This contribution was made on Mar 4
-
main: use jinja chat template system prompt by default
This contribution was made on Mar 1
Answered 1 discussion in 1 repository
ggml-org/llama.cpp
ggml-org/llama.cpp
-
How to use lazy grammars?
This contribution was made on Mar 6