Go
Home
Pricing
FAQ
Install
Home
Pricing
FAQ
Install
Login
via GitHub
ochafik/llama.cpp
Pull Requests
Commits
force-think
Telosnex_phi4_tools_template
Telosnex_phi4_tools_template2
add_ass_param
agent-example
agent-example2
agent-example4
agent-example5
agent-example6
agent-example-aug28
agent-example-cpp
agent-example-tmp
agent-example-tmp3
agent-examples6-jinjadev
allow-unclosed-think
bench-tool
bins
c7rb-fix
cache-url
chat-cleanups
chat-parser
ci-ccache
ci-fix-revocation-errors
ci-parallel
ci-write-less-ccache
claude/add-mmap-session-cache-vg6PE
claude/llamafile-build-script-dsKy5
cli
cli-tmp
cmake-b
cmake-b2
compact-bounds
cuda-releases
curl-fixes
date
decimal-bounds
detect-model-format
download
enable-thinking
enable-thinking2
enable-thinking-fix-chat-parsers
fast-quant
fix-chat-parsers
fix-diff-bug
fix-diff-bug2
fix-gen-prompt
fix-jinja-llg
fix-json-whitespace-parsing
fix-partial-regex-alternations
fix-python-tag
fix-qwen-tools
fix-reasoning-diffs
fix-stops
fn-name-delta
force-think
functiongemma-toolcall
generate-assets
generic-thoughts
gguf-read
gguf-read-field
grammar-async
grammar-async3
grammar-copy
grammar-fast
grammar-fast-disable
grammar-fast-linked
grammar-fast-nits
grammar-fast-resampling
grammar-fast-resampling-early
grammar-fast-single
grammar-fast-single2
grammar-from-json-schema
grammar-left-rec
grammar-less-stacks
grammar-rejects
grammar-reps
grammar-reps2
grammar-reps-revert
grammar-resample2
grammar-resampling
grammar-resampling-early
grammar-speedup
grammar-speedup2
grammar-speedup3
grammar-stacks-callbacks
grammar-transitions
grammars-early-exit
grammars-early-exit-resample
grammars-resample
gritlm-url
hip-ccache
install
jf-flag
jinja
jinja-chatml
json_schema
json-additional
json-bounds
json-doc
json-doc3
json-faster-repetitions
json-faster-repetitions2
json-faster-repetitions-bak
json-fixes
json-fixes-cpp
json-fixes-refact
json-integ
json-noadd
json-order
json-order2
json-order3
json-refs
json-refs2
json-schema-file
json-tests
json-type
json-updates
json-ws
kv-tree-cache
lazy-grammar-regex
llama2c-gguf
llama2c-non-shared-weights
llama2c-non-shared-weights-save
llama-lib-rename
lto2
main-json
master
mcp
minja-sync-aug8
minja-sync-minmax
model-args
nit-empty-grammar
ochafik/minja-sync2
offline
opencl-kernels
partial-regex
peg-end-tokens-metadata
peg-migration-complete
peg-migration-complete-history
peg-migration-inline-generic-helper
peg-migration-squashed
peg-migration-squashed-fixes
peg-simplify
quant-lowdisk
quant-lowraw-hack
quant-py
qwen3-chat
r1-toolcall
refector-chat-test
reuse-graphs
reuse-graphs-2
server-cancel
skip-unused
skip-unused-2-save
skip-unused-2
skip-unused-hipblas
stdin-prompt
string_utils
sync-minja-2
sync-minja-5
sync-minja-7
sync-minja-8
sync-minja-reasoning
sync-minja-reasoning-bak
tc-py
test-cli
tmp
tmp-grammar-debug-candidates
tmp-ochafik-masked-training
tmp-ochafik-opencl-clblast
tmp-ochafik-server-py
tmp-opencl-nofp16
token-aware-grammars-impl
tool-bench
tool-bench2
tool-bench-prod
tool-call
tool-call-0
tool-call-1
tool-call-2
tool-call-3
tool-call-bak1
tool-call-caching
tool-call-handler
tool-call-nits
tool-call-simpler
tool-diffs
tool-diffs-bak
tool-diffs-bak2
tool-diffs-bak3
update-functionary-tool-call-doc
url-meta
web-ui-mcp
Fix r1 grammar since we made <｜tool▁calls▁begin｜> optional (triggering on just <｜tool▁call▁begin｜> for 7B's sake)
ochafik
committed
1 year ago
0db98812
Merge branch 'sync-minja-4' into r1-toolcall
ochafik
committed
1 year ago
b5b117fa
ci : do not stale-close roadmap issues
ggerganov
committed
1 year ago
Verified
b34aedd5
Update chat.cpp
ochafik
committed
1 year ago
21f20715
fix test-chat
ochafik
committed
1 year ago
438ce0b8
ensure deepseek r1 thoughts parsed even w/o tool calls
ochafik
committed
1 year ago
1f5ec598
fix mistral expectation
ochafik
committed
1 year ago
b6e14a41
tool-call: ensure we don't return content when there are tool calls / warn
ochafik
committed
1 year ago
d44eb95c
server: check that content is null when we get tool_calls
ochafik
committed
1 year ago
812544ab
Merge branch 'sync-minja-4' into r1-toolcall
ochafik
committed
1 year ago
d43e4f6c
Update chat.cpp
ochafik
committed
1 year ago
f12e3507
fix mistral chat test: need empty tokens
ochafik
committed
1 year ago
56a14ddc
Update test-chat.cpp
ochafik
committed
1 year ago
b1527292
`sync`: minja
ochafik
committed
1 year ago
09caa634
fix spaces
ochafik
committed
1 year ago
86994db6
fix test_calc_result
ochafik
committed
1 year ago
78b47bb0
update test_calc_result
ochafik
committed
1 year ago
326e7002
Fix / test models/templates/llama-cpp-deepseek-r1.jinja
ochafik
committed
1 year ago
f0154a64
fix / test parsing of r1 parser
ochafik
committed
1 year ago
a682d121
move trigger_words init inside non-llguidance branch
ochafik
committed
1 year ago
9a6847c8
tool-call: r1: fix grammar
ochafik
committed
1 year ago
18a11f43
r1: fix inadvertent newline in grammar before <｜tool▁call▁end｜>
ochafik
committed
1 year ago
e84ee88f
tool-call: r1: add one more trigger approx "<｜tool calls begin｜>"
Olivier Chafik
committed
1 year ago
ce28224d
simplify hack to fix original template's backfill from minja
Olivier Chafik
committed
1 year ago
bff549de
sync: minja
Olivier Chafik
committed
1 year ago
bbd45bf6
update to minja's new api
Olivier Chafik
committed
1 year ago
30ea3591
actually we want eos_token in the template to infer tool call examples, explicitly skipped in new template options
Olivier Chafik
committed
1 year ago
11c1f0c7
Merge branch 'master' into r1-toolcall
Olivier Chafik
committed
1 year ago
bc6d910f
`tool-call`: allow `--chat-template chatml` w/ `--jinja`, default to chatml upon parsing issue, avoid double bos (#11616)
ochafik
committed
1 year ago
Verified
cde38332
sync: minja https://github.com/google/minja/pull/46
Olivier Chafik
committed
1 year ago
108da907
Newer
Older
ggerganov
committed 1 year ago
ochafik
committed 1 year ago