ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

mtmd : Support Pixtral 12B (#13065)

Browse source 
* add pixtral text model (vision is wip)

* cgraph ok, just missing 2D RoPE

* fix bad rebase

* first working version

* fix problem with img_break token

* support dynamic image size

* update docs

* update test script
This commit is contained in:
Xuan-Son Nguyen 2025-04-23 20:21:59 +02:00 • committed by GitHub
parent eb1776b15a
commit ecda2ec4b3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
14 changed files with 643 additions and 31 deletions
Download patch file Download diff file Expand all files Collapse all files

17
examples/llava/tests.sh
View file

@ -13,6 +13,14 @@ mkdir -p $SCRIPT_DIR/output
PROJ_ROOT="$SCRIPT_DIR/../.."
cd $PROJ_ROOT
# Check if the first argument is "big", then run test with big models
# This is useful if we're running the script on a larger machine, so we can test the big models
RUN_BIG_TESTS=false
if [ "${1:-}" = "big" ]; then
RUN_BIG_TESTS=true
echo "Include BIG models..."
fi
###############
arr_bin=()
@ -28,6 +36,12 @@ add_test() {
arr_tmpl+=("$tmpl")
}
add_test_big() {
if [ "$RUN_BIG_TESTS" = true ]; then
add_test "$@"
fi
}
add_test "llama-mtmd-cli" "ggml-org/SmolVLM-500M-Instruct-GGUF:Q8_0"
add_test "llama-mtmd-cli" "ggml-org/SmolVLM2-2.2B-Instruct-GGUF:Q4_K_M"
add_test "llama-mtmd-cli" "ggml-org/SmolVLM2-500M-Video-Instruct-GGUF:Q8_0"
@ -42,6 +56,9 @@ add_test "llama-mtmd-cli" "openbmb/MiniCPM-V-2_6-gguf:Q2_K"
add_test "llama-mtmd-cli" "openbmb/MiniCPM-o-2_6-gguf:Q4_0"
add_test "llama-qwen2vl-cli" "bartowski/Qwen2-VL-2B-Instruct-GGUF:Q4_K_M"
# to test the big models, run: ./tests.sh big
add_test_big "llama-mtmd-cli" "ggml-org/pixtral-12b-GGUF:Q4_K_M"
# these models always give the wrong answer, not sure why
# add_test "llama-mtmd-cli" "ggml-org/SmolVLM-Instruct-GGUF:Q4_K_M"
# add_test "llama-mtmd-cli" "ggml-org/SmolVLM-256M-Instruct-GGUF:Q8_0"