tests · llama-cpp-bonsai · Tangled

llama-cpp-bonsai

llama.cpp fork for Ternary Bonsai 2 27B: MTP speculation, suffix cache reuse (SCR), q8 KV, two-slot fairness, fused DeltaNet kernels

llama-cpp-bonsai tests

2 folders · 72 files

nekomimi.pet · cache: conversations are stored against their learned prefix · 1w ago

common: add json.h abstraction (#27511)

tests : avoid building get-model.cpp many times (#26317)

tests : add unit test coverage for llama_tensor_get_type (#20112)

cache: conversations are stored against their learned prefix

gen-tiny-dfly.py

dflash: fix DFly conversion + loading against the reference checkpoint

gen-tiny-dspark.py

dspark: add corrected draft runtime and shared packed target head

gguf-model-data.cpp

ci : add `[no release]` keyword + fix sanitizer builds (#23728)

gguf-model-data.h

tests : add unit test coverage for llama_tensor_get_type (#20112)

tests: Harmonize header use (#25616)

test-arg-parser.cpp

llama : support multi-output backend sampling (#25532)

test-autorelease.cpp

tests : avoid building get-model.cpp many times (#26317)

test-backend-ops.cpp

bonsai: in-place GDN recurrent state reading from cache rows (gdn-rows)

test-backend-sampler.cpp

tests : disable backend sampler hip multi output (#26878)

test-barrier.cpp

Fix race conditions in threadpool when dealing with dynamic/frequent n_threads changes (#17748)

test-batch-alloc.cpp

llama-batch: add unit test (#25471)

ggml : remove kompute backend (#14501)

test-chat-analysis.cpp

test: move tools/parser to tests (#27548)

test-chat-auto-parser.cpp

test: move tools/parser to tests (#27548)

test-chat-peg-parser.cpp

common: add json.h abstraction (#27511)

test-chat-template.cpp

test: move tools/parser to tests (#27548)

common: add json.h abstraction (#27511)

test-col2im-1d.cpp

ggml : add GGML_OP_COL2IM_1D (#24206)

test-dfly-fusion.cpp

dflash: AngelSpec DFly drafter support

test-double-float.cpp

ggml : minor naming changes (#8433)

test-dspark-forward.cpp

tests : build test-dspark-forward with MSVC (#273)

test-dspark-logsnr-meta.cpp

dspark: add corrected draft runtime and shared packed target head

test-dspark-loop.cpp

dspark: validate mask selection and graph reuse contracts

test-dspark-real-eval.cpp

dspark: validate mask selection and graph reuse contracts

test-export-graph-ops.cpp

tests: export-graph-ops: exit gracefully when called w/o arguments (#25619)

test-gbnf-validator.cpp

cmake : do not include ./src as public for libllama (#13062)

test-gguf-model-data.cpp

tests : add unit test coverage for llama_tensor_get_type (#20112)

gguf : harden loader against malformed tensor dims and metadata types (#25596)

test-grammar-integration.cpp

common: add json.h abstraction (#27511)

test-grammar-llguidance.cpp

tool/ex/tests: consistently free ctx, then model (#18168)

test-grammar-parser.cpp

grammar : degrade max repetition >= 2000 to unbounded (#26613)

common: add json.h abstraction (#27511)

test-json-schema-to-grammar.cpp

common: add json.h abstraction (#27511)

test-kv-mean-center.cpp

kv-cache: mean-centering on hybrid models + calibration-basis guard (fork #53)

test-kv-share.cpp

cache: conversations are stored against their learned prefix

test-llama-archs.cpp

tests : disable DOTS3NOTE arch test for WebGPU (#27654)

test-llama-grammar.cpp

common/grammar: fix grammar parsing issues to prevent stack overflow and hangs (#18604)

common: Intentionally leak logger instance to fix hanging on Windows (#22273)

test-lora-conversion-inference.sh

cli: new CLI experience (#17824)

test-model-load-cancel.cpp

tests : avoid building get-model.cpp many times (#26317)

test-model-resolution.cpp

common: add json.h abstraction (#27511)

test-mtmd-c-api.c

mtmd: add chunk save/load function (#26645)

test-mtmd-impl.cpp

mtmd: add mtmd_bitmap_set_mergeable (#27348)

tests : fix test-opt with GGML_BACKEND_DL (#15599)

test-peg-parser.cpp

Autoparser - complete refactoring of parser architecture (#18675)

test-pq2-row-shapes.cpp

ggml-cpu: keep PQ2_0 rows that are not multiples of 256 on the Q8_0 activation path

test-ptq1_0-cuda-dot.cpp

rocm: vectorized HIP PTQ1_0 vec_dot (#211)

test-ptq1_0-element-map.cpp

ggml: add PTQ1_0, ternary at group 128 (1.75 bpw, lossless vs PQ2_0) (#148)

test-quant-type-selection.cpp

tests : avoid building get-model.cpp many times (#26317)

test-quantize-fns.cpp

x86: SSE2/SSSE3 vec_dot for PTQ1_0 and PQ2_0 (#248)

test-quantize-perf.cpp

ci: run the x64 and arm ci on the github machines instead (#16183)

test-quantize-stats.cpp

cmake : introduce semantic versioning (#26839)

test-reasoning-budget.cpp

common : add support for multiple end sequences in the reasoning budget sampler (#25544)

test-recurrent-state-rollback.cpp

DeepseekV4: fix rollback with multi-seq (#26756)

ggml-cpu: templateify ggml_compute_forward_rope_f32 and _f16 (#16805)

test-rset-release.cpp

tests : avoid building get-model.cpp many times (#26317)

test-sampling.cpp

llama : support multi-output backend sampling (#25532)

test-save-load-state.cpp

tests : remove unnecessary sync in test-save-load-state (#26166)

scr: retain up to 64 matching spans for fragmented transcripts

test-state-restore-fragmented.cpp

common : only load backends when required (#22290)

test-thread-safety.cpp

tests : synchronize contexts at end of test-thread-safety (#24935)

test-tokenizer-0.cpp

tool/ex/tests: consistently free ctx, then model (#18168)

test-tokenizer-0.py

requirements : update transformers to 5.5.1 (#21617)

test-tokenizer-0.sh

model : add Jina Embeddings v5 Nano (partial EuroBERT) support (#19826)

test-tokenizer-1-bpe.cpp

tool/ex/tests: consistently free ctx, then model (#18168)

test-tokenizer-1-spm.cpp

tool/ex/tests: consistently free ctx, then model (#18168)

test-tokenizer-random.py

requirements : update transformers to 5.5.1 (#21617)

test-tokenizers-repo.sh

devops: add s390x & ppc64le CI (#15925)

test-unicode.cpp

unicode : include '~' in collapsed symbol class (#26972)

chat: refactor handling supports_string_content / supports_typed_content (#27130)