repo stringclasses 15
values | fix_commit stringlengths 40 40 | buggy_commit stringlengths 40 40 | message stringlengths 3 64.3k | files listlengths 1 300 | timestamp timestamp[s]date 2013-03-13 20:45:00 2026-04-11 07:48:46 |
|---|---|---|---|---|---|
ollama/ollama | 013fd071395087019b4e05959d8249f28ac11930 | f63dc2db5c00de0f6b0b5ea9b53bb20e83513cea | More logging for gpu management
Fix an ordering glitch of dlerr/dlclose and add more logging to help
root cause some crashes users are hitting. This also refines the
function pointer names to use the underlying function names instead
of simplified names for readability. | [
{
"path": "gpu/gpu.go",
"patch": "@@ -40,11 +40,13 @@ var CudaLinuxGlobs = []string{\n \t\"/usr/lib/wsl/lib/libnvidia-ml.so*\",\n \t\"/usr/lib/wsl/drivers/*/libnvidia-ml.so*\",\n \t\"/opt/cuda/lib64/libnvidia-ml.so*\",\n-\t\"/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so*\",\n \t\"/usr/lib*/libnvi... | 2024-01-24T18:32:00 |
ggml-org/llama.cpp | 2ba9adc093127016a48cd0c5d6bf1420dafe17a6 | cc45f2ada695644c6697c0fb0e70a5e95563ad0f | Adjust workaround for ROCWMMA_FATTN/GFX9 to only newer ROCm veresions (#19591)
Avoids issues with ROCm 6.4.4.
Closes: https://github.com/ggml-org/llama.cpp/issues/19580
Fixes: 6845f7f87 ("Add a workaround for compilation with ROCWMMA_FATTN and gfx9 (#19461)")
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.or... | [
{
"path": "ggml/src/ggml-cuda/fattn-wmma-f16.cu",
"patch": "@@ -63,7 +63,7 @@ static __global__ void flash_attn_ext_f16(\n constexpr int frag_m = ncols == 8 ? 32 : 16;\n constexpr int frag_n = ncols == 8 ? 8 : 16;\n static_assert(D % frag_m == 0, \"If ncols == 8 then D % frag_m must be 0.\");\n... | 2026-02-16T13:46:08 |
rust-lang/rust | 6efacfb7a59ebde2620398861713fae136060a04 | 7279acf2025a5affd51862d126624ed0f49f701b | add fix for full tools and sanitizer | [
{
"path": "Cargo.lock",
"patch": "@@ -1997,22 +1997,22 @@ dependencies = [\n ]\n \n [[package]]\n-name = \"libffi\"\n-version = \"3.2.0\"\n+name = \"libffi-sys2\"\n+version = \"2.4.0\"\n source = \"registry+https://github.com/rust-lang/crates.io-index\"\n-checksum = \"ce826c243048e3d5cec441799724de52e2d42f8... | 2025-02-24T03:42:07 |
huggingface/transformers | 04c7c176d7f70ec4b43c8c2a0327ff8d193f5c1d | 42d8dd8716d45291733191464e249a83b93e4d86 | [tests] make `test_model_parallelism` device-agnostic (#30844)
* enable on xpu
* fix style
* add comment and mps | [
{
"path": "tests/test_modeling_common.py",
"patch": "@@ -76,6 +76,7 @@\n require_safetensors,\n require_torch,\n require_torch_gpu,\n+ require_torch_multi_accelerator,\n require_torch_multi_gpu,\n require_torch_sdpa,\n slow,\n@@ -3009,8 +3010,11 @@ def check_device_map_is_respecte... | 2024-05-24T10:51:51 |
golang/go | 076a45acd5b8d2ce08a2dbe898dc9228554db92d | e08059f4fcacce2ff18d1cfc5fa48c942f8d46aa | fmt: explain how Formatter interface affects verbs and flags
Formatter is mentioned further down, but it's helpful
to add it amongst the verbs and flags.
Background: I spent a while puzzling how "%+v" prints
a stack trace for github.com/pkg/errors when this isn't
documented under 'flags'.
Change-Id: Ic70145902a36780... | [
{
"path": "src/fmt/doc.go",
"patch": "@@ -150,7 +150,8 @@\n \tconcrete value that it holds, and printing continues with the next rule.\n \n \t2. If an operand implements the Formatter interface, it will\n-\tbe invoked. Formatter provides fine control of formatting.\n+\tbe invoked. In this case the interpret... | 2020-10-12T14:29:28 |
nodejs/node | 75095d700137e29ad1413972b8f6eed7c3473a65 | b6df87e1d482f18202586273200fd7a8e9bc2fcd | test: mark test-async-wrap-uncaughtexception as flaky
`parallel/test-async-wrap-uncaughtexception.js` has become flaky.
At this time investigating the cause is still on going, but this issue
become has prevalent. In order to restore CI status to be relevant,
this marks the test as explicitly FLAKY.
PR-URL: https://gi... | [
{
"path": "test/parallel/parallel.status",
"patch": "@@ -5,6 +5,8 @@ prefix parallel\n # sample-test : PASS,FLAKY\n \n [true] # This section applies to all platforms\n+# https://github.com/nodejs/node/issues/16210\n+test-async-wrap-uncaughtexception\n \n [$system==win32]\n ",
"add... | 2017-11-02T17:07:05 |
ggml-org/llama.cpp | d5dfc330279a76fa25614db1c6413599e8974da3 | 267ba5a1d957158316a4fc85b1f0cf316d9a5233 | graph : fix KQ mask, lora, cvec reuse checks (#19644)
* graph : fix KQ mask reuse condition
* cont : dedup KQ mask build and can_reuse
* cont : fix build
* graph : fix adapter check for reuse | [
{
"path": "src/llama-adapter.h",
"patch": "@@ -39,6 +39,8 @@ struct llama_adapter_cvec {\n std::vector<ggml_tensor *> tensors; // per layer\n };\n \n+using llama_adapter_cvec_ptr = std::shared_ptr<llama_adapter_cvec>;\n+\n //\n // llama_adapter_lora\n //\n@@ -84,3 +86,4 @@ struct llama_adapter_lora {\n ... | 2026-02-16T07:21:11 |
ollama/ollama | 987c16b2f755c0ac7f3c3b03d4b875dffc96a551 | 0759d8996e10277f2879d99cca03fa2174d0ac3c | Report more information about GPUs in verbose mode
This adds additional calls to both CUDA and ROCm management libraries to
discover additional attributes about the GPU(s) detected in the system, and
wires up runtime verbosity selection. When users hit problems with GPUs we can
ask them to run with `OLLAMA_DEBUG=1 ol... | [
{
"path": "gpu/gpu.go",
"patch": "@@ -259,6 +259,7 @@ func FindGPULibs(baseLibName string, patterns []string) []string {\n \n func LoadCUDAMgmt(cudaLibPaths []string) *C.cuda_handle_t {\n \tvar resp C.cuda_init_resp_t\n+\tresp.ch.verbose = getVerboseState()\n \tfor _, libPath := range cudaLibPaths {\n \t\tl... | 2024-01-23T00:03:32 |
rust-lang/rust | e6a4cf63adfaed64b74ed04441e902ce64a47116 | a9c61ec1e1445d4eb8a8eabf0bb858ac220bd345 | don't trigger `unnecessary_debug_formatting` in tests | [
{
"path": "clippy_lints/src/format_args.rs",
"patch": "@@ -9,7 +9,7 @@ use clippy_utils::macros::{\n use clippy_utils::msrvs::{self, Msrv};\n use clippy_utils::source::{SpanRangeExt, snippet};\n use clippy_utils::ty::{implements_trait, is_type_lang_item};\n-use clippy_utils::{is_diag_trait_item, is_from_pro... | 2025-03-04T01:50:40 |
huggingface/transformers | 42d8dd8716d45291733191464e249a83b93e4d86 | 5855afd1f38e8b263710a04b7619662716428a36 | Perceiver interpolate position embedding (#30979)
* add test that currently fails
* test passed
* all perceiver passed
* fixup, style, quality, repo-consistency, all passed
* Apply suggestions from code review: default to False + compute sqrt once only
Co-authored-by: amyeroberts <22614925+amyeroberts@u... | [
{
"path": "src/transformers/models/perceiver/modeling_perceiver.py",
"patch": "@@ -699,13 +699,24 @@ def _init_weights(self, module):\n output_hidden_states (`bool`, *optional*):\n Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for\n ... | 2024-05-24T10:13:58 |
golang/go | ad53103aef09c002c41ea34292cfea359857ae5b | 58e51b1e620167cc22ca7143c395cb63db5640a8 | io: add ErrBadWriteCount
Fixes #39978
Change-Id: Ib41459861ba9f7cf0bf1fc95b1479c358c4bdbd8
GitHub-Last-Rev: 19cbb1461ca04a8eb64f0c4f354d8fb81a70d4f3
GitHub-Pull-Request: golang/go#39989
Reviewed-on: https://go-review.googlesource.com/c/go/+/240740
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Go Bot <g... | [
{
"path": "src/io/io.go",
"patch": "@@ -30,6 +30,9 @@ var ErrShortWrite = errors.New(\"short write\")\n // ErrShortBuffer means that a read required a longer buffer than was provided.\n var ErrShortBuffer = errors.New(\"short buffer\")\n \n+// ErrBadWriteCount means that a write returned an impossible count... | 2020-10-13T05:56:48 |
ggml-org/llama.cpp | 27b93cbd157fc4ad94573a1fbc226d3e18ea1bb4 | 6e67fd2144e1647bf88b027c0807b5de35705acc | cuda: optimize iq2xxs/iq2xs/iq3xxs dequantization (#19624)
* cuda: optimize iq2xxs/iq2xs/iq3xxs dequantization
- load all 8 int8 for a grid position in one load
- calculate signs via popcnt instead of fetching from ksigns table
- broadcast signs to drop individual shift/mask
* cuda: iq2xxs: simplify sum scaling
exp... | [
{
"path": "ggml/src/ggml-cuda/mmq.cuh",
"patch": "@@ -2715,14 +2715,14 @@ template <int mmq_y, bool need_check> static __device__ __forceinline__ void loa\n \n #pragma unroll\n for (int l = 0; l < QR2_XXS; ++l) {\n- const int * grid_pos = (const int *) (iq2xxs_grid + aux8[l]);\n- ... | 2026-02-15T17:08:42 |
nodejs/node | b6df87e1d482f18202586273200fd7a8e9bc2fcd | 14181a3368b38978e3178d66a549349cc720f707 | http, tls: better support for IPv6 addresses
- Properly handle IPv6 in Host header when setting servername.
- When comparing IP addresses against addresses in the subjectAltName
field of a certificate, format the address correctly before
doing the string comparison.
PR-URL: https://github.com/nodejs/node/pull/147... | [
{
"path": "lib/_http_agent.js",
"patch": "@@ -153,13 +153,8 @@ Agent.prototype.addRequest = function addRequest(req, options, port/*legacy*/,\n if (options.socketPath)\n options.path = options.socketPath;\n \n- if (!options.servername) {\n- options.servername = options.host;\n- const hostHeader... | 2017-08-10T13:11:19 |
ollama/ollama | 0f5b8433195c505a673b5687897bfa7379ab3d13 | ffaf52e1e9e107c2b6946a6593b780f941d0a337 | Refine Accelerate usage on mac
For old macs, accelerate seems to cause crashes, but for
AVX2 capable macs, it does not. | [
{
"path": "llm/generate/gen_darwin.sh",
"patch": "@@ -18,7 +18,7 @@ sign() {\n fi\n }\n \n-COMMON_DARWIN_DEFS=\"-DCMAKE_OSX_DEPLOYMENT_TARGET=11.0 -DCMAKE_SYSTEM_NAME=Darwin -DLLAMA_ACCELERATE=off\"\n+COMMON_DARWIN_DEFS=\"-DCMAKE_OSX_DEPLOYMENT_TARGET=11.0 -DCMAKE_SYSTEM_NAME=Darwin\"\n \n case \"${GOAR... | 2024-01-23T00:25:56 |
huggingface/transformers | 5855afd1f38e8b263710a04b7619662716428a36 | 03935d300d60110bb86edb49d2315089cfb19789 | pin `uv==0.1.45` (#31006)
* fix
* [push-ci-image]
* run with latest
---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com> | [
{
"path": "docker/consistency.dockerfile",
"patch": "@@ -4,7 +4,7 @@ USER root\n ARG REF=main\n RUN apt-get update && apt-get install -y time git pkg-config make git-lfs\n ENV VIRTUAL_ENV=/usr/local\n-RUN pip install uv && uv venv && uv pip install --no-cache-dir -U pip setuptools GitPython\n+RUN pip instal... | 2024-05-24T10:00:50 |
ggml-org/llama.cpp | 9e118b97c456d378dea49a8c46b850c71fc18363 | 57088276d42faf7a0f4a347578cff2c17159906d | build : remove LLAMA_HTTPLIB option (#19623)
This option was introduced as a workaround because cpp-httplib could not
build on visionOS. Since it has been fixed and now compiles on all platforms,
we can remove it and simplify many things.
Signed-off-by: Adrien Gallouët <angt@huggingface.co> | [
{
"path": "CMakeLists.txt",
"patch": "@@ -112,7 +112,6 @@ option(LLAMA_TOOLS_INSTALL \"llama: install tools\" ${LLAMA_TOOLS_INSTALL_\n option(LLAMA_TESTS_INSTALL \"llama: install tests\" ON)\n \n # 3rd party libs\n-option(LLAMA_HTTPLIB \"llama: httplib for downloading functionality\" ON)\... | 2026-02-15T14:38:50 |
golang/go | 58e51b1e620167cc22ca7143c395cb63db5640a8 | a15ec3895b99cd6b49576a6c1a739b46f8afeaf4 | cmd/go/internal/fsys: skip symlink test on Plan 9
Fixes #41950
Fixes #41954
Change-Id: I95d97f076fa928f3638309b78748d7ccc7277b14
Reviewed-on: https://go-review.googlesource.com/c/go/+/261897
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Michael Matloob <matloob@... | [
{
"path": "src/cmd/go/internal/fsys/fsys_test.go",
"patch": "@@ -5,6 +5,7 @@ import (\n \t\"encoding/json\"\n \t\"errors\"\n \t\"fmt\"\n+\t\"internal/testenv\"\n \t\"io/ioutil\"\n \t\"os\"\n \t\"path/filepath\"\n@@ -654,6 +655,8 @@ func TestWalk_Error(t *testing.T) {\n }\n \n func TestWalk_Symlink(t *testin... | 2020-10-13T15:36:11 |
rust-lang/rust | 4febd273e5a82a9c5aa648e8d843baa184c30929 | fd17deacce374a4185c882795be162e17b557050 | Fix some typos
Signed-off-by: fuyangpengqi <995764973@qq.com> | [
{
"path": "CONTRIBUTING.md",
"patch": "@@ -18,7 +18,7 @@ For submodules, changes need to be made against the repository corresponding the\n submodule, and not the main `rust-lang/rust` repository.\n \n For subtrees, prefer sending a PR against the subtree's repository if it does\n-not need to be made agains... | 2025-03-04T08:05:32 |
nodejs/node | e22a8f17a5d70355062b483337692e45954380a0 | 89b3228f4d1a7f00b8ff195964895061e45861d1 | src: improve module loader readability
Various improvements on readability, performance and
conformity to the Node core coding style in the ESM loader C++ code:
- `isolate` for the `Isolate*`, `context` for the `Local<Context>`
- Less reliance on `auto` where it’s unnecessary/increases cognitive
overhead
- Shorter ... | [
{
"path": "src/env.h",
"patch": "@@ -185,6 +185,7 @@ class ModuleWrap;\n V(kill_signal_string, \"killSignal\") \\\n V(length_string, \"length\") \\\n V(mac_string, \"mac\") ... | 2017-10-26T23:43:51 |
huggingface/transformers | 21e259d8c578759bc1062d2537f97e766f78cfa9 | 7f6e87413f03b41b0896b372e826b27df08ed4c8 | Fix training speed regression introduced by "optimize VRAM for calculating pos_bias in LayoutLM v2, v3 (#26139)" (#30988)
* Revert "optimize VRAM for calculating pos_bias in LayoutLM v2, v3 (#26139)"
This reverts commit a7e0ed829c398a67a641a401e23dae13e2f8b217.
* Instead of reverting commit, wrap indexing in tor... | [
{
"path": "src/transformers/models/layoutlmv2/modeling_layoutlmv2.py",
"patch": "@@ -383,7 +383,12 @@ def _calculate_1d_position_embeddings(self, position_ids):\n num_buckets=self.rel_pos_bins,\n max_distance=self.max_rel_pos,\n )\n- rel_pos = self.rel_pos_bias.weight.... | 2024-05-24T08:43:44 |
ggml-org/llama.cpp | 57088276d42faf7a0f4a347578cff2c17159906d | 341bc7d23cf8bd77639f0c5b9dcb2dc8e9713768 | cmake : check if KleidiAI API has been fetched (#19640)
This commit addresses a build issue with the KleidiAI backend when
building multiple cpu backends. Commmit
3a00c98584e42a20675b6569d81beadb282b0952 ("cmake : fix KleidiAI install
target failure with EXCLUDE_FROM_ALL") introduced a change where
FetchContent_Popula... | [
{
"path": "ggml/src/ggml-cpu/CMakeLists.txt",
"patch": "@@ -576,13 +576,13 @@ function(ggml_add_cpu_backend_variant_impl tag_name)\n DOWNLOAD_EXTRACT_TIMESTAMP NEW\n URL_HASH MD5=${KLEIDIAI_ARCHIVE_MD5})\n \n- FetchContent_Populate(KleidiAI_Download)\n FetchContent_Get... | 2026-02-15T12:59:38 |
golang/go | a15ec3895b99cd6b49576a6c1a739b46f8afeaf4 | 575ea5a87a86de3107b2401e10ff4e50e5a133a0 | cmd/link: update plt size appropriately on ppc64
When attempting to enable internal linking with cgo on ppc64
it was discovered that the plt size was not being updated
after adding entries to it, which resulted in this error:
.plt: initialize bounds (16 < 24)
This changes fixes that problem.
Updates #21961
Chang... | [
{
"path": "src/cmd/link/internal/ppc64/asm.go",
"patch": "@@ -994,6 +994,7 @@ func addpltsym(ctxt *ld.Link, ldr *loader.Loader, s loader.Sym) {\n \t\tldr.SetPlt(s, int32(plt.Size()))\n \n \t\tplt.Grow(plt.Size() + 8)\n+\t\tplt.SetSize(plt.Size() + 8)\n \n \t\trela.AddAddrPlus(ctxt.Arch, plt.Sym(), int64(ldr... | 2020-10-13T13:15:52 |
nodejs/node | 3a977fc8c0d13f796c65e05e31bb548217fbc37b | 9ad994befbb319e31b2fef3551134ff95be06069 | test: fix flaky test-http2-server-rst-stream.js
PR-URL: https://github.com/nodejs/node/pull/16690
Fixes: https://github.com/nodejs/node/issues/16688
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com> | [
{
"path": "test/parallel/test-http2-server-rst-stream.js",
"patch": "@@ -5,75 +5,52 @@ if (!common.hasCrypto)\n common.skip('missing crypto');\n const assert = require('assert');\n const http2 = require('http2');\n+const Countdown = require('../common/countdown');\n \n const {\n- HTTP2_HEADER_METHOD,\n- ... | 2017-11-02T16:46:31 |
huggingface/transformers | 7f6e87413f03b41b0896b372e826b27df08ed4c8 | 6657fb5fed3f94812ef97dda16e91eee77f03d61 | add prefix space ignored in llama #29625 (#30964)
* add prefix space ignored in llama #29625
* adding test with add_prefix_space=False
* ruff
---------
Co-authored-by: Ita Zaporozhets <itazaporozhets@Itas-MBP.localdomain> | [
{
"path": "src/transformers/models/llama/tokenization_llama_fast.py",
"patch": "@@ -163,6 +163,7 @@ def __init__(\n add_bos_token=add_bos_token,\n add_eos_token=add_eos_token,\n use_default_system_prompt=use_default_system_prompt,\n+ add_prefix_space=add_prefix... | 2024-05-24T08:03:00 |
ggml-org/llama.cpp | 341bc7d23cf8bd77639f0c5b9dcb2dc8e9713768 | 08e6d914b8fef477d2a5aeeb89e08d42709981cf | context : fix output reorder with backend sampling (#19638) | [
{
"path": "src/llama-context.cpp",
"patch": "@@ -878,6 +878,7 @@ const llama_token * llama_context::get_sampled_candidates_ith(int32_t idx) {\n }\n } catch (const std::exception & err) {\n // fallback to full vocab list\n+ GGML_UNUSED(err);\n }\n \n return sampling.token_i... | 2026-02-15T12:57:40 |
golang/go | 575ea5a87a86de3107b2401e10ff4e50e5a133a0 | 5b509d993d3a3a212b4033815be8b7b439fac672 | cmd/link: set runtime.text to the address of the first function
In CL 240065 we changed it to set to FlagTextAddr. Normally it
is the address of the first function, except on plan9/amd64
where, as FlagTextAddr is not aligned, it is rounded up. Set it
to the actual text start address.
Fixes #41137.
Change-Id: I1bba67... | [
{
"path": "src/cmd/link/internal/ld/data.go",
"patch": "@@ -2188,7 +2188,7 @@ func (ctxt *Link) textaddress() {\n \t\tctxt.Textp[0] = text\n \t}\n \n-\tva := uint64(*FlagTextAddr)\n+\tva := uint64(Rnd(*FlagTextAddr, int64(Funcalign)))\n \tn := 1\n \tsect.Vaddr = va\n \tntramps := 0\n@@ -2214,7 +2214,7 @@ fu... | 2020-10-12T22:00:25 |
huggingface/transformers | 6657fb5fed3f94812ef97dda16e91eee77f03d61 | 6d3d5b1039559a7cc03965cc8c2a6ff0291d83fc | Bugfix: WandbCallback uploads initial model checkpoint (#30897)
* fix wandb always uploading initial model
* Update comment.
* Optionally log initial model
* Revert "Optionally log initial model"
This reverts commit 9602cc1fad3feaf218f82a7339a194d3d2fbb946. | [
{
"path": "src/transformers/integrations/integration_utils.py",
"patch": "@@ -797,7 +797,7 @@ def setup(self, args, state, model, **kwargs):\n except AttributeError:\n logger.info(\"Could not log the number of model parameters in Weights & Biases.\")\n \n- # log the in... | 2024-05-23T19:29:00 |
ollama/ollama | 730dcfcc7ab3c1fc24417887533eee8c72e45a1c | 5f81a33f43edea71edfb3d045e140595caeaa226 | Refine debug logging for llm
This wires up logging in llama.cpp to always go to stderr, and also
turns up logging if OLLAMA_DEBUG is set. | [
{
"path": "llm/dyn_ext_server.go",
"patch": "@@ -136,6 +136,12 @@ func newDynExtServer(library, model string, adapters, projectors []string, opts\n \n \tsparams.n_threads = C.uint(opts.NumThread)\n \n+\tif debug := os.Getenv(\"OLLAMA_DEBUG\"); debug != \"\" {\n+\t\tsparams.verbose_logging = C.bool(true)\n+\... | 2024-01-22T20:26:49 |
nodejs/node | 9ad994befbb319e31b2fef3551134ff95be06069 | 3d9d84940ae11a026c8f14e1b502794db7538765 | dgram: migrate bufferSize to use internal/errors
PR-URL: https://github.com/nodejs/node/pull/16567
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com> | [
{
"path": "lib/dgram.js",
"patch": "@@ -173,11 +173,13 @@ function bufferSize(self, size, buffer) {\n if (size >>> 0 !== size)\n throw new errors.TypeError('ERR_SOCKET_BAD_BUFFER_SIZE');\n \n- try {\n- return self._handle.bufferSize(size, buffer);\n- } catch (e) {\n- throw new errors.Error('ER... | 2017-10-28T06:24:15 |
ggml-org/llama.cpp | 684b36101c9eeb7e89c9e602f9ded05f1353a0c6 | 3a00c98584e42a20675b6569d81beadb282b0952 | ggml-cpu: FA add GEMM microkernel (#19422)
* ggml-cpu: FA add GEMM microkernel
* add guard for sizeless vector types
* fix case where DV % GGML_F32_EPR !=0
* move memset out of the loop
* move another memset out of the loop
* use RM=4 for arm
* simd_gemm: convert everything to int
* convert everything to size_t... | [
{
"path": "ggml/src/ggml-cpu/common.h",
"patch": "@@ -6,8 +6,8 @@\n #include \"ggml-impl.h\"\n #include \"simd-mappings.h\"\n \n-#define GGML_FA_TILE_Q 32\n-#define GGML_FA_TILE_KV 16\n+#define GGML_FA_TILE_Q 64\n+#define GGML_FA_TILE_KV 64\n \n #ifdef __cplusplus\n ",
"additions": 2,
"deletions":... | 2026-02-15T05:39:24 |
golang/go | e69f6e839315b1df4e6273c38ae1a49e340b8a91 | c8fdfa756ead0a48b0227f6cec02641797124a92 | internal/bytealg: fix typo in IndexRabinKarp{,Bytes} godoc
Change-Id: I09ba19e19b195e345a0fe29d542e0d86529b0d31
Reviewed-on: https://go-review.googlesource.com/c/go/+/261359
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
R... | [
{
"path": "src/internal/bytealg/bytealg.go",
"patch": "@@ -99,7 +99,7 @@ func HashStrRev(sep string) (uint32, uint32) {\n }\n \n // IndexRabinKarpBytes uses the Rabin-Karp search algorithm to return the index of the\n-// first occurence of substr in s, or -1 if not present.\n+// first occurrence of substr i... | 2020-10-12T09:19:41 |
huggingface/transformers | 6d3d5b1039559a7cc03965cc8c2a6ff0291d83fc | 965e98dc54c4853b34b3a49810f24856a9fcdba8 | Remove deprecated properties in tokenization_nllb.py and tokenization_nllb_fast.py (#29834)
* Fix typo in tokenization_nllb.py
Change `adder_tokens_decoder` into `added_tokens_decoder` and improve the warning's readability.
* Fix typo in tokenization_nllb_fast.py
Change `adder_tokens_decoder` into `added_toke... | [
{
"path": "src/transformers/models/nllb/tokenization_nllb.py",
"patch": "@@ -159,18 +159,6 @@ def __init__(\n self.fairseq_offset = 1\n self.sp_model_size = len(self.sp_model)\n \n- # Everything that follows is kept for BC and will be removed in v4.38\n- self._fairseq_tokens_to... | 2024-05-23T16:53:26 |
nodejs/node | 3d9d84940ae11a026c8f14e1b502794db7538765 | 056b858e57a72c5429d9278bc7b77d1922a2dbe4 | tty: convert to internal/errors using SystemError
PR-URL: https://github.com/nodejs/node/pull/16567
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com> | [
{
"path": "lib/internal/errors.js",
"patch": "@@ -18,6 +18,7 @@ const { defineProperty } = Object;\n \n // Lazily loaded\n var util = null;\n+var buffer;\n \n function makeNodeError(Base) {\n return class NodeError extends Base {\n@@ -59,6 +60,12 @@ function makeNodeError(Base) {\n };\n }\n \n+function ... | 2017-10-28T00:20:54 |
ollama/ollama | 27a2d5af5452b82a5d1ab44853fd55a2c63865be | 5f81a33f43edea71edfb3d045e140595caeaa226 | Debug logging on init failure | [
{
"path": "llm/dyn_ext_server.go",
"patch": "@@ -142,7 +142,9 @@ func newDynExtServer(library, model string, adapters, projectors []string, opts\n \tC.dyn_llama_server_init(llm.s, &sparams, &initResp)\n \tif initResp.id < 0 {\n \t\tmutex.Unlock()\n-\t\treturn nil, extServerResponseToErr(initResp)\n+\t\terr ... | 2024-01-22T20:08:22 |
ggml-org/llama.cpp | 3a00c98584e42a20675b6569d81beadb282b0952 | 079feab9e3efee1d6d4ca370eac50f156e2dc6e8 | cmake : fix KleidiAI install target failure with EXCLUDE_FROM_ALL (#19581)
* cmake: fix KleidiAI install target failure with EXCLUDE_FROM_ALL
Fix for the bug #19501 by adding EXCLUDE_FROM_ALL to FetchContent_Declare. This properly excludes KleidiAI from both build and install targets, preventing install failures when... | [
{
"path": "ggml/src/ggml-cpu/CMakeLists.txt",
"patch": "@@ -569,12 +569,14 @@ function(ggml_add_cpu_backend_variant_impl tag_name)\n cmake_policy(SET CMP0135 NEW)\n endif()\n \n+ # TODO: Use FetchContent_MakeAvailable with EXCLUDE_FROM_ALL after bumping minimum CMake version to 3.... | 2026-02-15T05:22:53 |
golang/go | c8fdfa756ead0a48b0227f6cec02641797124a92 | 85f829deb8adfb7c0d73acabfdb458de969f6763 | go/build: reject //go:build without // +build
We are converting from using error-prone ad-hoc syntax // +build lines
to less error-prone, standard boolean syntax //go:build lines.
The timeline is:
Go 1.16: prepare for transition
- Builds still use // +build for file selection.
- Source files may not contain //go:bu... | [
{
"path": "src/go/build/build.go",
"patch": "@@ -1371,9 +1371,12 @@ func (ctxt *Context) matchFile(dir, name string, allTags map[string]bool, binary\n \t}\n \n \t// Look for +build comments to accept or reject the file.\n-\tok, sawBinaryOnly := ctxt.shouldBuild(data, allTags)\n+\tok, sawBinaryOnly, err := c... | 2020-05-22T18:46:52 |
huggingface/transformers | 965e98dc54c4853b34b3a49810f24856a9fcdba8 | 2a89673fe5b946097cd483964abe543a7868aa9b | [Port] TensorFlow implementation of Mistral (#29708)
* chore: initial commit
* chore: adding imports and inits
* chore: adding the causal and classification code
* chore: adding names to the layers
* chore: using single self attn layer
* chore: built the model and layers
* chore: start with testing
... | [
{
"path": "docs/source/en/index.md",
"patch": "@@ -200,7 +200,7 @@ Flax), PyTorch, and/or TensorFlow.\n | [Megatron-BERT](model_doc/megatron-bert) | ✅ | ❌ | ❌ |\n | [Megatron-GPT2](model_doc/megatron_gpt2) ... | 2024-05-23T16:48:49 |
nodejs/node | 056b858e57a72c5429d9278bc7b77d1922a2dbe4 | c3dc0e0d75cc5a03de154f3586e1062df1e719e4 | os: migrate node_os.cc to internal/errors
PR-URL: https://github.com/nodejs/node/pull/16567
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com> | [
{
"path": "lib/os.js",
"patch": "@@ -27,21 +27,43 @@ const { deprecate } = require('internal/util');\n const { getCIDRSuffix } = require('internal/os');\n const isWindows = process.platform === 'win32';\n \n+const errors = require('internal/errors');\n+\n const {\n getCPUs,\n getFreeMem,\n- getHomeDire... | 2017-10-27T23:27:16 |
ggml-org/llama.cpp | 01d8eaa28d57bfc6d06e30072085ed0ef12e06c5 | 1725e316c1a780759ec134ca5a2999f4d53ce273 | mtmd : Add Nemotron Nano 12B v2 VL support (#19547)
* nemotron nano v2 vlm support added
* simplified code; addressed reviews
* pre-downsample position embeddings during GGUF conversion for fixed input size | [
{
"path": "convert_hf_to_gguf.py",
"patch": "@@ -4074,6 +4074,87 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter\n yield from super().modify_tensors(data_torch, name, bid)\n \n \n+@ModelBase.register(\n+ \"NemotronH_Nano_VL_V2\",\n+ \"RADIOModel\",\n... | 2026-02-14T13:07:00 |
golang/go | 85f829deb8adfb7c0d73acabfdb458de969f6763 | 2c6df2e35dd95621d025306a49b863efd77ebf3a | cmd/asm: reject misplaced go:build comments
We are converting from using error-prone ad-hoc syntax // +build lines
to less error-prone, standard boolean syntax //go:build lines.
The timeline is:
Go 1.16: prepare for transition
- Builds still use // +build for file selection.
- Source files may not contain //go:buil... | [
{
"path": "src/cmd/asm/internal/asm/endtoend_test.go",
"patch": "@@ -257,11 +257,11 @@ func isHexes(s string) bool {\n \treturn true\n }\n \n-// It would be nice if the error messages began with\n+// It would be nice if the error messages always began with\n // the standard file:line: prefix,\n // but that'... | 2020-06-29T21:08:49 |
ggml-org/llama.cpp | 1725e316c1a780759ec134ca5a2999f4d53ce273 | b7742cf3217932b2e237861c8586b6f600f072fb | models : optimize qwen3next graph (#19375)
* models : optimizing qwen3next graph
* cont
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* cont : remove redundant q, g chunking
* minor
* minor
* avoid passing masks around
* avoid concats during chunking
* naming + shapes
* update names and... | [
{
"path": "ggml/src/ggml-cuda/ggml-cuda.cu",
"patch": "@@ -2872,6 +2872,7 @@ static bool ggml_cuda_graph_check_compability(ggml_cgraph * cgraph) {\n const std::string ffn_moe_down_bias_prefix = \"ffn_moe_down_biased\";\n const std::string nemotron_h_block_out_prefix = \"nemotron_h_block_out\";\n ... | 2026-02-14T10:57:36 |
ollama/ollama | ec3764538d4755448b326400093608040219fd5f | fa8c990e58c05f100d728420a33feb45400223d8 | Probe GPUs before backend init
Detect potential error scenarios so we can fallback to CPU mode without
hitting asserts. | [
{
"path": "llm/ext_server/ext_server.cpp",
"patch": "@@ -3,6 +3,27 @@\n // Necessary evil since the server types are not defined in a header\n #include \"server.cpp\"\n \n+// Low level API access to verify GPU access\n+#if defined(GGML_USE_CUBLAS)\n+#if defined(GGML_USE_HIPBLAS)\n+#include <hip/hip_runtime.... | 2024-01-21T23:39:59 |
nodejs/node | c3dc0e0d75cc5a03de154f3586e1062df1e719e4 | f1f0eb25951a33cd97661aee44a9b43bc6eb6d27 | src: add CollectExceptionInfo & errors.SystemError
Preparing for the migration of existing UVException and ErrnoExceptions
from the native layer, add new `errors.SystemError` to internal/errors
and new `env->CollectExceptionInfo()` / `env->CollectUVExceptionInfo()`
methods.
PR-URL: https://github.com/nodejs/node/pull... | [
{
"path": "doc/api/errors.md",
"patch": "@@ -458,7 +458,7 @@ checks or `abort()` calls in the C++ layer.\n \n ## System Errors\n \n-System errors are generated when exceptions occur within the program's\n+System errors are generated when exceptions occur within the Node.js\n runtime environment. Typically, ... | 2017-10-27T23:25:14 |
huggingface/transformers | 2a89673fe5b946097cd483964abe543a7868aa9b | 892b13d3cf592bda9a6506ed526ba5b6c80bbefc | Update 4 `MptIntegrationTests` expected outputs (#30989)
* fix
* fix
* fix
* fix
* fix
* [run-slow] mpt
---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com> | [
{
"path": "tests/models/mpt/test_modeling_mpt.py",
"patch": "@@ -447,7 +447,7 @@ def test_generation_8k(self):\n )\n \n input_text = \"Hello\"\n- expected_output = 'Hello, I\\'m a new user of the forum. I have a question about the \"Safety\"'\n+ expected_output = \"Hello, I'm a... | 2024-05-23T16:27:54 |
ggml-org/llama.cpp | b7742cf3217932b2e237861c8586b6f600f072fb | badba89320e9ba8cbebe15a7dfd33983512b76c3 | ggml : fix GGML_DEBUG with OpenMP (#19599)
last_graph is only available without OpenMP, but
ggml_graph_compute_thread() is called in both cases.
Signed-off-by: Adrien Gallouët <angt@huggingface.co> | [
{
"path": "ggml/src/ggml-cpu/ggml-cpu.c",
"patch": "@@ -2947,7 +2947,11 @@ static thread_ret_t ggml_graph_compute_thread(void * data) {\n /*.use_ref =*/ cplan->use_ref,\n };\n \n- GGML_PRINT_DEBUG(\"thread #%d compute-start cplan %p last-graph %d \\n\", state->ith, cplan, state->last_graph... | 2026-02-14T10:22:57 |
golang/go | 2c6df2e35dd95621d025306a49b863efd77ebf3a | ec095f1df0ff1b2937205e17b3c011f31d3737ca | cmd/compile: reject misplaced go:build comments
We are converting from using error-prone ad-hoc syntax // +build lines
to less error-prone, standard boolean syntax //go:build lines.
The timeline is:
Go 1.16: prepare for transition
- Builds still use // +build for file selection.
- Source files may not contain //go:... | [
{
"path": "src/cmd/compile/internal/gc/lex.go",
"patch": "@@ -50,6 +50,9 @@ const (\n \n \t// Runtime and cgo type pragmas\n \tNotInHeap // values of this type must not be heap allocated\n+\n+\t// Go command pragmas\n+\tGoBuildPragma\n )\n \n const (\n@@ -71,6 +74,8 @@ const (\n \n func pragmaFlag(verb stri... | 2020-05-22T19:22:52 |
ollama/ollama | c08dfaa23d90d5df044f44f1ba89fc1e0b727757 | 3b76e736aed78e19879c184b6d44570e0a150c36 | fix: remove overwritten model layers
if create overrides a manifest, first add the older manifest's layers to
the delete map so they can be cleaned up | [
{
"path": "server/images.go",
"patch": "@@ -412,6 +412,13 @@ func realpath(mfDir, from string) string {\n }\n \n func CreateModel(ctx context.Context, name, modelFileDir string, commands []parser.Command, fn func(resp api.ProgressResponse)) error {\n+\tdeleteMap := make(map[string]struct{})\n+\tif manifest,... | 2024-01-19T22:58:36 |
rust-lang/rust | 09e584671b439dc8a9d57cc900cc26668a69ddea | 06072468fe8de249a39d2fcc42cea51db2649a80 | Also note struct access, and fix macro expansion from foreign crates | [
{
"path": "compiler/rustc_hir_analysis/src/hir_ty_lowering/errors.rs",
"patch": "@@ -153,7 +153,7 @@ impl<'tcx> dyn HirTyLowerer<'tcx> + '_ {\n sugg: None,\n // Try to get the span of the identifier within the path's syntax context\n // (if that's different).\n- ... | 2025-02-24T21:37:01 |
nodejs/node | f1f0eb25951a33cd97661aee44a9b43bc6eb6d27 | d4471e06e8424035c214369d8a4480bb4c9e4f69 | test: pause child until parent is ready
Pause child on startup using inspect-brk=0 until the parent debugger
is ready.
PR-URL: https://github.com/nodejs/node/pull/15774
Fixes: https://github.com/nodejs/node/issues/14897
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com... | [
{
"path": "test/sequential/sequential.status",
"patch": "@@ -10,7 +10,6 @@ prefix sequential\n test-inspector-async-call-stack : PASS, FLAKY\n test-inspector-bindings : PASS, FLAKY\n test-inspector-debug-end : PASS, FLAKY\n-test-inspector-stop-profile-after-done: PASS, FLAKY... | 2017-10-04T05:18:00 |
huggingface/transformers | 892b13d3cf592bda9a6506ed526ba5b6c80bbefc | 21339a52133ba62ddc2ba5d12c68a6c053f7a4df | Add a check that warmup_setps is either 0 or >= 1 (#30764)
* Add a check that warmup_setps is either 0 or >= 1
Update training_args.py to add a check that warmup_setps is either 0 or >= 1. Otherwise, raise an error.
* Update src/transformers/training_args.py
Co-authored-by: amyeroberts <22614925+amyeroberts@u... | [
{
"path": "src/transformers/training_args.py",
"patch": "@@ -1797,6 +1797,9 @@ def __post_init__(self):\n \" during training\"\n )\n \n+ if not isinstance(self.warmup_steps, int) or self.warmup_steps < 0 or 0 < self.warmup_steps <= 1:\n+ raise ValueError(\"warmu... | 2024-05-23T16:23:59 |
rust-lang/rust | 06072468fe8de249a39d2fcc42cea51db2649a80 | e4dfca8ac240d517e5d7008333fd7d3367d0176f | Fix associated type errors too | [
{
"path": "compiler/rustc_hir_analysis/messages.ftl",
"patch": "@@ -620,6 +620,8 @@ hir_analysis_variances_of = {$variances}\n hir_analysis_where_clause_on_main = `main` function is not allowed to have a `where` clause\n .label = `main` cannot have a `where` clause\n \n+hir_analysis_within_macro = due t... | 2025-02-24T20:36:15 |
golang/go | 5ef78c4d84620f6594ad63cbd1b3b43f41670636 | 7bda6154caa6f0c527f4a8302e38d450b44ae68b | hash/crc32: fix race between lazy Castagnoli init and Update/Write
The switch on tab is checking tab == castagnoliTable,
but castagnoliTable can change value during a concurrent
call to MakeTable.
Fixes #41911.
Change-Id: I6124dcdbf33e17fe302baa3e1aa03202dec61b4c
Reviewed-on: https://go-review.googlesource.com/c/go/... | [
{
"path": "src/hash/crc32/crc32.go",
"patch": "@@ -16,6 +16,7 @@ import (\n \t\"errors\"\n \t\"hash\"\n \t\"sync\"\n+\t\"sync/atomic\"\n )\n \n // The size of a CRC-32 checksum in bytes.\n@@ -78,6 +79,7 @@ var castagnoliTable8 *slicing8Table\n var castagnoliArchImpl bool\n var updateCastagnoli func(crc uint... | 2020-10-12T19:41:14 |
huggingface/transformers | 21339a52133ba62ddc2ba5d12c68a6c053f7a4df | 8366b5724103b10e1dcd8755d2ed6defff435fbb | [tests] add `torch.use_deterministic_algorithms` for XPU (#30774)
* add xpu check
* add marker
* add documentation
* update doc
* fix ci
* remove from global init
* fix | [
{
"path": "src/transformers/testing_utils.py",
"patch": "@@ -116,6 +116,7 @@\n is_torch_bf16_available_on_device,\n is_torch_bf16_cpu_available,\n is_torch_bf16_gpu_available,\n+ is_torch_deterministic,\n is_torch_fp16_available_on_device,\n is_torch_neuroncore_available,\n is_tor... | 2024-05-23T15:53:07 |
nodejs/node | 841e305e4c0bc9f633006a293b893fd69dbbf596 | 367db920e3b69ed9ded50cf6e652b9f581e10748 | doc : mention constant-time in crypto doc
Included reference to 'constant time' in
crypto.timingSafeEqual description.
The Node website would score higher on a Google search
and the API would be more discoverable if it used the words
"constant time" in its description.
PR-URL: https://github.com/nodejs/node/pull/166... | [
{
"path": "doc/api/crypto.md",
"patch": "@@ -1969,6 +1969,7 @@ added: v6.6.0\n - `a` {Buffer | TypedArray | DataView}\n - `b` {Buffer | TypedArray | DataView}\n \n+This function is based on a constant-time algorithm.\n Returns true if `a` is equal to `b`, without leaking timing information that\n would allo... | 2017-10-30T15:52:18 |
rust-lang/rust | e4dfca8ac240d517e5d7008333fd7d3367d0176f | ab31129956869f6dfc25417230061bad72de9081 | Point out macro expansion ident in resolver errors too | [
{
"path": "compiler/rustc_hir_typeck/src/method/suggest.rs",
"patch": "@@ -194,7 +194,8 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {\n node => unreachable!(\"{node:?}\"),\n };\n \n- // Try to get the span of the identifier within the expression's syntax context (if that's different).\n+ ... | 2025-02-24T20:20:54 |
golang/go | ca3c0df1f8e07337ba4048b191bf905118ebe251 | c321430bdc328c394de501bdc5f8f5d6d8952cd4 | net/http/httputil: flush ReverseProxy immediately if Content-Length is -1
Finish up a prior TODO by making ReverseProxy flush immediately
if Content-Length is -1, which is a case that can occur if for
example we have a streamed response, or chunked encoding, or when
the body's length wasn't known.
Fixes #41642
Chang... | [
{
"path": "src/net/http/httputil/reverseproxy.go",
"patch": "@@ -58,9 +58,9 @@ type ReverseProxy struct {\n \t// A negative value means to flush immediately\n \t// after each write to the client.\n \t// The FlushInterval is ignored when ReverseProxy\n-\t// recognizes a response as a streaming response;\n-\t... | 2020-10-08T20:32:50 |
nodejs/node | 67c8511ea13b6def8c39042b10749a9b182feeac | 94b2be7a96a5b50022ea523af069c96dfb559fe0 | src: use internal/errors for startSigintWatchdog
Move the throw out of c++ and into js using internal/errors
PR-URL: https://github.com/nodejs/node/pull/16546
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed... | [
{
"path": "doc/api/errors.md",
"patch": "@@ -615,6 +615,11 @@ Used when attempting to perform an operation outside the bounds of a `Buffer`.\n Used when an attempt has been made to create a `Buffer` larger than the\n maximum allowed size.\n \n+<a id=\"ERR_CANNOT_WATCH_SIGINT\"></a>\n+### ERR_CANNOT_WATCH_SI... | 2017-10-27T05:04:07 |
huggingface/transformers | 8366b5724103b10e1dcd8755d2ed6defff435fbb | 5a74ae6dbe84da6017546ebd3765da6cd08dbc40 | Fix accelerate failing tests (#30836)
* Fix accelerate tests
* fix clip
* skip dbrx tests
* fix GPTSan
* fix M2M100Model
* same fix as jamba
* fix mt5
* Fix T5Model
* Fix umt5 model
* fix switch_transformers
* fix whisper
* fix gptsan again
* fix siglip recent test
* skip siglip ... | [
{
"path": "src/transformers/models/bert/modeling_bert.py",
"patch": "@@ -964,7 +964,7 @@ class BertModel(BertPreTrainedModel):\n `add_cross_attention` set to `True`; an `encoder_hidden_states` is then expected as an input to the forward pass.\n \"\"\"\n \n- _no_split_modules = [\"BertEmbeddings\"... | 2024-05-23T15:18:58 |
golang/go | 112c4d569e8881cc9910039bf4c45176a9b1f619 | 2f4368c174f616b6623e66ca4c84e701d209fa4f | cmd/fix: print warning to stderr, show provenance
This cgo warning shows up when running "go test" in cmd/fix:
warning: no cgo types: exit status 1
(It may show up at other times too.)
Warnings belong on standard error, and while we're here,
make it say where it came from.
Change-Id: I3eaba592a219cc6abf2d58476237... | [
{
"path": "src/cmd/fix/typecheck.go",
"patch": "@@ -207,7 +207,7 @@ func typecheck(cfg *TypeConfig, f *ast.File) (typeof map[interface{}]string, ass\n \t\t\treturn nil\n \t\t}()\n \t\tif err != nil {\n-\t\t\tfmt.Printf(\"warning: no cgo types: %s\\n\", err)\n+\t\t\tfmt.Fprintf(os.Stderr, \"go fix: warning: ... | 2020-06-29T18:20:04 |
ollama/ollama | eb76f3e37993ab0c7d1c0d1c361245e1597690aa | d017e3d0a6a95922f9efe2625824e708a3db13f9 | Fix CPU-only build under Android Termux enviornment.
Update gpu.go initGPUHandles() to declare gpuHandles variable before
reading it. This resolves an "invalid memory address or nil pointer
dereference" error.
Update dyn_ext_server.c to avoid setting the RTLD_DEEPBIND flag under
__TERMUX__ (Android). | [
{
"path": "gpu/gpu.go",
"patch": "@@ -63,6 +63,7 @@ func initGPUHandles() {\n \n \t// TODO - if the ollama build is CPU only, don't do these checks as they're irrelevant and confusing\n \n+\tgpuHandles = &handles{nil, nil}\n \tvar cudaMgmtName string\n \tvar cudaMgmtPatterns []string\n \tvar rocmMgmtName st... | 2024-01-15T09:37:44 |
nodejs/node | 94b2be7a96a5b50022ea523af069c96dfb559fe0 | 78dbcbe559324cdd73c2a158ad6ce77a7a74717f | doc: fix Changelog link order
PR-URL: https://github.com/nodejs/node/pull/16632
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Anatoli Papirovski <apapirovski@mac.com>
Reviewed-By: Tobias Nießen <tniessen@tni... | [
{
"path": "CHANGELOG.md",
"patch": "@@ -10,11 +10,12 @@ Select a Node.js version below to view the changelog history:\n * [Node.js 8](doc/changelogs/CHANGELOG_V8.md)\n * [Node.js 7](doc/changelogs/CHANGELOG_V7.md)\n * [Node.js 6](doc/changelogs/CHANGELOG_V6.md)\n+* [Node.js 5](doc/changelogs/CHANGELOG_V5.md... | 2017-10-31T01:44:59 |
ggml-org/llama.cpp | 2d8015e8a460f51a4c0fc8b1b9f41e38d7ea2194 | eb145c0753e702d74447f427f7233d519ad1f8e6 | llama : update LoRA API. + fix excessive graph reserves (#19280)
* Refactoring to use new llama_put_adapter_loras
* cont : alternative lora API
---------
Co-authored-by: Jake Chavis <jakechavis6@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> | [
{
"path": "common/common.cpp",
"patch": "@@ -1223,7 +1223,7 @@ common_init_result_ptr common_init_from_params(common_params & params) {\n return res;\n }\n \n- int err = llama_apply_adapter_cvec(\n+ int err = llama_set_adapter_cvec(\n lctx,\n ... | 2026-02-14T08:06:27 |
huggingface/transformers | 5a74ae6dbe84da6017546ebd3765da6cd08dbc40 | 046c2ad7923bb0601f53d6f32c3092aa697ab14d | FIX / Docs: Minor changes in quantization docs (#30985)
* Change in quantization docs
* Update overview.md
* Update docs/source/en/quantization/overview.md
Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
---------
Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com> | [
{
"path": "docs/source/en/quantization/overview.md",
"patch": "@@ -24,14 +24,14 @@ Interested in adding a new quantization method to Transformers? Read the [HfQuan\n \n </Tip>\n \n-> [!TIP]\n-> If you are new to the quantization field, we recommend you to check out these beginner-friendly courses about quan... | 2024-05-23T14:36:49 |
golang/go | 2f4368c174f616b6623e66ca4c84e701d209fa4f | 09833da6b4cfb494ba61a398ab2063ab016d3890 | cmd/fix: rename confusing boolean
("truth" says nothing about what is true.
This boolean tracks whether the file was fixed.)
Change-Id: I29bb80c4fad3ca7f2ae96e50e16f6cde484b374f
Reviewed-on: https://go-review.googlesource.com/c/go/+/240556
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com... | [
{
"path": "src/cmd/fix/gotypes.go",
"patch": "@@ -21,11 +21,11 @@ var gotypesFix = fix{\n }\n \n func gotypes(f *ast.File) bool {\n-\ttruth := fixGoTypes(f)\n+\tfixed := fixGoTypes(f)\n \tif fixGoExact(f) {\n-\t\ttruth = true\n+\t\tfixed = true\n \t}\n-\treturn truth\n+\treturn fixed\n }\n \n func fixGoType... | 2020-06-29T18:10:24 |
ggml-org/llama.cpp | eb145c0753e702d74447f427f7233d519ad1f8e6 | 6e473fb384c21126ba3f1d5b5ce81ccbb17d465a | mmap: Fix Windows handle lifetime (#19598)
* ggml: added cleanups in ggml_quantize_free
Add missing cleanup calls for IQ2_S, IQ1_M quantization types and IQ3XS with 512 blocks during quantization cleanup.
* mmap: Fix Windows handle lifetime
Move hMapping from local variable to member variable so it stays alive for th... | [
{
"path": "src/llama-mmap.cpp",
"patch": "@@ -504,14 +504,16 @@ struct llama_mmap::impl {\n }\n }\n #elif defined(_WIN32)\n+ HANDLE hMapping = nullptr;\n+\n impl(struct llama_file * file, size_t prefetch, bool numa) {\n GGML_UNUSED(numa);\n \n size = file->size();\n \n ... | 2026-02-14T08:05:12 |
nodejs/node | c52fe67162862e96ba9d7b600f7b6bbe9ba74efd | 7489ee84fe992fb6e8c1d77e9b8801f5d783cdfe | test: use fixtures module in test-https-pfx
PR-URL: https://github.com/nodejs/node/pull/15895
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com> | [
{
"path": "test/parallel/test-https-pfx.js",
"patch": "@@ -21,14 +21,16 @@\n \n 'use strict';\n const common = require('../common');\n+\n if (!common.hasCrypto)\n common.skip('missing crypto');\n \n+const fixtures = require('../common/fixtures');\n+\n const assert = require('assert');\n-const fs = require... | 2017-10-06T17:11:23 |
golang/go | 09833da6b4cfb494ba61a398ab2063ab016d3890 | 7b77ff4c88befb1da9fb8b1cc7128a44ffb8a64f | cmd/fix: always format source file before fixing
This makes the changes to the file easier to explain.
Not all the changes may come from the fixers directly,
if the file is not gofmt-ed already.
Change-Id: I81776da446a34a1239a3130317d2aae1437d61a5
Reviewed-on: https://go-review.googlesource.com/c/go/+/240555
Trust: R... | [
{
"path": "src/cmd/fix/main.go",
"patch": "@@ -137,6 +137,21 @@ func processFile(filename string, useStdin bool) error {\n \t\treturn err\n \t}\n \n+\t// Make sure file is in canonical format.\n+\t// This \"fmt\" pseudo-fix cannot be disabled.\n+\tnewSrc, err := gofmtFile(file)\n+\tif err != nil {\n+\t\tret... | 2020-06-29T14:44:56 |
huggingface/transformers | 87a351818e6c680282d86a00b61fe4c55b41477d | d583f1317be422128b7bb56984720387b3bbcb35 | Docs / Quantization: refactor quantization documentation (#30942)
* refactor quant docs
* delete file
* rename to overview
* fix
* fix table
* fix
* add content
* fix library versions
* fix table
* fix table
* fix table
* fix table
* Apply suggestions from code review
Co-authored... | [
{
"path": "docs/source/en/_toctree.yml",
"patch": "@@ -135,18 +135,36 @@\n title: Community resources\n - local: troubleshooting\n title: Troubleshoot\n- - local: hf_quantizer\n- title: Contribute new quantization method\n - local: gguf\n title: Interoperability with GGUF files\n title... | 2024-05-23T12:31:52 |
ggml-org/llama.cpp | 6e473fb384c21126ba3f1d5b5ce81ccbb17d465a | c7db95f10638550392deaa85e65d6df7e886e7ab | metal : fix ACC op (#19427) | [
{
"path": "ggml/src/ggml-metal/ggml-metal-device.m",
"patch": "@@ -1067,8 +1067,8 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te\n case GGML_OP_MUL:\n case GGML_OP_DIV:\n case GGML_OP_ADD_ID:\n- return ggml_is_contiguous_rows(op->src[0]... | 2026-02-14T07:54:03 |
rust-lang/rust | c566318a782030a33f370d7490d7bdac9d8bfca4 | e16a049adbf94d610787430b6efdf31d896dc5b6 | Tweak error code for sized checks of const/static | [
{
"path": "compiler/rustc_hir_analysis/src/check/wfcheck.rs",
"patch": "@@ -1108,7 +1108,13 @@ fn check_associated_item(\n let ty = tcx.type_of(item.def_id).instantiate_identity();\n let ty = wfcx.normalize(span, Some(WellFormedLoc::Ty(item_id)), ty);\n wfcx.r... | 2025-02-18T17:11:12 |
golang/go | e0cb80e4d10bb9967d70ff1a0d7e66c9d8af21cf | aeae46a7e51921b6af1dc1400bdc341fc1e568b0 | cmd/compile/internal/syntax: provide BadExpr where needed, call correct error handler
- For "if" statements without a condition, provide a BadExpr rather than nil
(clients expect IfStmt.Cond != nil since the parser is taking care of
reporting a missing condition).
- For 3-index slice expressions, also provide Bad... | [
{
"path": "src/cmd/compile/internal/syntax/parser.go",
"patch": "@@ -287,6 +287,7 @@ func tokstring(tok token) string {\n \n // Convenience methods using the current token position.\n func (p *parser) pos() Pos { return p.posAt(p.line, p.col) }\n+func (p *parser) error(msg string) { p.er... | 2020-10-10T00:42:41 |
ggml-org/llama.cpp | c7db95f10638550392deaa85e65d6df7e886e7ab | 0d00ef65ed3fe5e995247ef9b9f83d5467b9191e | scripts : use official split.py for cpp-httplib (#19588)
* scripts : use official split.py for cpp-httplib
Using the official script is safer and ensures the generated code aligns
with the library's standards.
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
* Catch generic errors
Signed-off-by: Adrien Gallouë... | [
{
"path": "scripts/sync_vendor.py",
"patch": "@@ -1,6 +1,9 @@\n #!/usr/bin/env python3\n \n import urllib.request\n+import os\n+import sys\n+import subprocess\n \n HTTPLIB_VERSION = \"f80864ca031932351abef49b74097c67f14719c6\"\n \n@@ -14,7 +17,8 @@\n # \"https://github.com/mackron/miniaudio/raw/refs/tag... | 2026-02-14T07:41:16 |
huggingface/transformers | d583f1317be422128b7bb56984720387b3bbcb35 | e05baad86119efcbf1e2dae64eef0c29f589b07b | Quantized KV Cache (#30483)
* clean-up
* Update src/transformers/cache_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/cache_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/cache_utils.p... | [
{
"path": "docker/transformers-all-latest-gpu/Dockerfile",
"patch": "@@ -51,6 +51,9 @@ RUN python3 -m pip install --no-cache-dir gguf\n # Some slow tests require bnb\n RUN python3 -m pip install --no-cache-dir bitsandbytes\n \n+# Some tests require quanto\n+RUN python3 -m pip install --no-cache-dir quanto\n... | 2024-05-23T12:25:20 |
ollama/ollama | aac9ab4db756b999c6c95b1159fc357a3b7c44b3 | 1f5b7ff976460ad5305ab31998554944825f1f5f | fix show handler | [
{
"path": "server/routes.go",
"patch": "@@ -630,11 +630,10 @@ func ShowModelHandler(c *gin.Context) {\n \t\treturn\n \t}\n \n-\tvar model string\n \tif req.Model != \"\" {\n-\t\tmodel = req.Model\n+\t\t// noop\n \t} else if req.Name != \"\" {\n-\t\tmodel = req.Name\n+\t\treq.Model = req.Name\n \t} else {\n ... | 2024-01-18T23:36:50 |
golang/go | aeae46a7e51921b6af1dc1400bdc341fc1e568b0 | 1aa43a53be83c3eaa468cae4914a55878fb2dad9 | cmd/go: disable automatic go vet -unreachable during go test of std
go test runs a limited number of vet checks by default.
In the standard library, we run more, both to get additional checking
for the standard library and to get experience with whether to enable
any others by default.
One that experience has shown u... | [
{
"path": "src/cmd/go/internal/work/exec.go",
"patch": "@@ -1052,17 +1052,28 @@ func (b *Builder) vet(ctx context.Context, a *Action) error {\n \t// This is OK as long as the packages that are farther down the\n \t// dependency tree turn on *more* analysis, as here.\n \t// (The unsafeptr check does not writ... | 2020-06-22T18:45:35 |
huggingface/transformers | 4ef85fee718969f1703d7dffa134deb72f4de828 | eb1a77bbb0a62d721e9a02e67b7e4f9e5afca08b | Push ci image (#30982)
* [build-ci-image]
* correct branch
* push ci image
* [build-ci-image]
* update scheduled as well
* [push-ci-image]
* [build-ci-image]
* [push-ci-image]
* update deps
* [build-ci-image]
* [build-ci-image]
* [build-ci-image]
* [build-ci-image]
* [build-ci-im... | [
{
"path": ".circleci/create_circleci_config.py",
"patch": "@@ -75,7 +75,7 @@ def __post_init__(self):\n else:\n # BIG HACK WILL REMOVE ONCE FETCHER IS UPDATED\n print(os.environ.get(\"GIT_COMMIT_MESSAGE\"))\n- if \"[build-ci-image]\" in os.environ.get(\"GIT_COMMIT_... | 2024-05-23T09:45:31 |
ggml-org/llama.cpp | 91ea5d67f2a47fff99592c46583758131bfcdf62 | dbb023336bc6986f3d1bc292502d7badea3713c5 | build : fix libtool call in build-xcframework.sh (#19605)
Run libtool via xcrun like strip and dsymutil, to have proper tool resolution.
Signed-off-by: Adrien Gallouët <angt@huggingface.co> | [
{
"path": "build-xcframework.sh",
"patch": "@@ -43,11 +43,6 @@ COMMON_CMAKE_ARGS=(\n -DGGML_OPENMP=${GGML_OPENMP}\n )\n \n-XCODE_VERSION=$(xcodebuild -version 2>/dev/null | head -n1 | awk '{ print $2 }')\n-MAJOR_VERSION=$(echo $XCODE_VERSION | cut -d. -f1)\n-MINOR_VERSION=$(echo $XCODE_VERSION | cut -d.... | 2026-02-14T05:48:37 |
golang/go | 9384d34c58099657bb1b133beaf3ff37ada9b017 | 39b527691495902279da7ac8405a070ded7dd4a2 | text/template: allow newlines inside action delimiters
This allows multiline constructs like:
{{"hello" |
printf}}
Now that unclosed actions can span multiple lines,
track and report the start of the action when reporting errors.
Also clean up a few "unexpected <error message>" to be just "<error message>".
Fi... | [
{
"path": "src/html/template/exec_test.go",
"patch": "@@ -1302,7 +1302,7 @@ func TestUnterminatedStringError(t *testing.T) {\n \t\tt.Fatal(\"expected error\")\n \t}\n \tstr := err.Error()\n-\tif !strings.Contains(str, \"X:3: unexpected unterminated raw quoted string\") {\n+\tif !strings.Contains(str, \"X:3:... | 2020-09-10T22:53:26 |
huggingface/transformers | eb1a77bbb0a62d721e9a02e67b7e4f9e5afca08b | 15585b81a525e3878321df4b13404c96690a2469 | Using assistant in AutomaticSpeechRecognitionPipeline with different encoder size (#30637)
* fiw input to generate in pipeline
* fixup
* pass input_features to generate with assistant
* error if model and assistant with different enc size
* fix
* apply review suggestions
* use self.config.is_encoder_... | [
{
"path": "src/transformers/generation/utils.py",
"patch": "@@ -1097,6 +1097,25 @@ def _validate_model_class(self):\n exception_message += f\" Please use one of the following classes instead: {generate_compatible_classes}\"\n raise TypeError(exception_message)\n \n+ def _valid... | 2024-05-23T08:59:38 |
nodejs/node | 8cd940af13e1718be57ddfbfeb07c634091b4981 | e509db8c564a660231cc09afef3d900f58201afc | 2017-10-31, Version 9.0.0 (Current)
* Older experimental APIs have been removed.
[[`d731369b1d`](https://github.com/nodejs/node/commit/d731369b1d)]
[#14414](https://github.com/nodejs/node/pull/14414)
* **Errors**
* Improvements have been made to `buffer` module error messages.
* The assignment of static... | [
{
"path": "CHANGELOG.md",
"patch": "@@ -6,6 +6,7 @@ release lines.\n \n Select a Node.js version below to view the changelog history:\n \n+* [Node.js 9](doc/changelogs/CHANGELOG_V9.md)\n * [Node.js 8](doc/changelogs/CHANGELOG_V8.md)\n * [Node.js 7](doc/changelogs/CHANGELOG_V7.md)\n * [Node.js 6](doc/changel... | 2017-09-01T16:50:47 |
ggml-org/llama.cpp | 0ccbfdef3e0a635530aec490f863d83edc22cbc4 | 94a602db6693fa6d2529d284fc04a8e9107e20de | hexagon: further optimizations and refactoring for flash attention (#19583)
* ggml-hexagon: fa improvements
ggml-hexagon: optimize flash attention calculations with improved variable handling
ggml-hexagon: streamline flash attention operations by removing redundant checks for FP32
ggml-hexagon: optimize hvx_dot_f16... | [
{
"path": "ggml/src/ggml-hexagon/htp/flash-attn-ops.c",
"patch": "@@ -17,121 +17,6 @@\n #include \"htp-msg.h\"\n #include \"htp-ops.h\"\n \n-static inline HVX_Vector hvx_load_f32_to_f16(const HVX_Vector * restrict src, const HVX_Vector zero) {\n- HVX_Vector y0_qf = Q6_Vqf32_vsub_VsfVsf(src[0], zero); //... | 2026-02-14T00:27:30 |
rust-lang/rust | 69ef2fc908f79d97c6deeee3e2a220688782a62d | d4916623403bc29927f467ccb1a80ba836a04139 | fix: properly escape regexes | [
{
"path": "src/etc/lldb_batchmode.py",
"patch": "@@ -40,7 +40,7 @@ def print_debug(s):\n \n def normalize_whitespace(s):\n \"\"\"Replace newlines, tabs, multiple spaces, etc with exactly one space\"\"\"\n- return re.sub(\"\\s+\", \" \", s)\n+ return re.sub(r\"\\s+\", \" \", s)\n \n \n def breakpoi... | 2025-03-03T21:22:26 |
ollama/ollama | 96cfb626415cd811ab545c134bd0d16fa7aca044 | 598d6d5572337afac759abf1995950866d96d044 | fix: normalize name path before splitting | [
{
"path": "server/modelpath.go",
"patch": "@@ -46,6 +46,7 @@ func ParseModelPath(name string) ModelPath {\n \t\tname = after\n \t}\n \n+\tname = strings.ReplaceAll(name, string(os.PathSeparator), \"/\")\n \tparts := strings.Split(name, \"/\")\n \tswitch len(parts) {\n \tcase 3:",
"additions": 1,
"de... | 2024-01-17T00:48:05 |
huggingface/transformers | 15585b81a525e3878321df4b13404c96690a2469 | a25f7d3c12975fe21eab437dda7363e9024de7c0 | Update object detection with latest resize and pad strategies (#30955)
* Update with new resizing and pad strategy
* Return pixel mask param
* Update inference in guide
* Fix empty compose
* Update guide | [
{
"path": "docs/source/en/tasks/object_detection.md",
"patch": "@@ -206,49 +206,39 @@ Instantiate the image processor from the same checkpoint as the model you want t\n \n >>> image_processor = AutoImageProcessor.from_pretrained(\n ... MODEL_NAME,\n-... # At this moment we recommend using external t... | 2024-05-22T23:13:56 |
golang/go | 0f53453b32054b681dac6e8b025c85a785396102 | e92ce92662172196ebe5099de93ee5bdabaa3f18 | net/http: deep copy Request.TransferEncoding
The existing implementation in Request.Clone() assigns the wrong
pointer to r2.TransferEncoding.
Fixes #41907
Change-Id: I7f220a41b1b46a55d1a1005e47c6dd69478cb025
Reviewed-on: https://go-review.googlesource.com/c/go/+/261258
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.... | [
{
"path": "src/net/http/request.go",
"patch": "@@ -382,7 +382,7 @@ func (r *Request) Clone(ctx context.Context) *Request {\n \tif s := r.TransferEncoding; s != nil {\n \t\ts2 := make([]string, len(s))\n \t\tcopy(s2, s)\n-\t\tr2.TransferEncoding = s\n+\t\tr2.TransferEncoding = s2\n \t}\n \tr2.Form = cloneURL... | 2020-10-10T20:25:07 |
ggml-org/llama.cpp | 752584d5f5f30bff8f24a765f858645a9ef7d5d9 | cc2aa81513ff8bbf8f0527ce36b9366971d5fd35 | model: support GLM MoE DSA arch (NOTE: indexer is not yet supported) (#19460)
* model: support GLM MoE DSA arch
* working version
* pyright
* keep indexer tensors
* add indexer gguf params
* loaded now
* Apply suggestions from code review
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
* update
... | [
{
"path": "convert_hf_to_gguf.py",
"patch": "@@ -1608,6 +1608,23 @@ def _set_vocab_glmedge(self):\n special_vocab._set_special_token(\"bos\", tokenizer.get_added_vocab()[\"<|endoftext|>\"])\n special_vocab.add_to_gguf(self.gguf_writer)\n \n+ def _set_vocab_glm(self):\n+ from transf... | 2026-02-13T13:56:53 |
huggingface/transformers | a25f7d3c12975fe21eab437dda7363e9024de7c0 | d44e1ae03646617b550ff628a706ffd8fe093767 | Paligemma causal attention mask (#30967)
* PaliGemma working causal attention
* Formatting
* Style
* Docstrings + remove commented code
* Update docstring for PaliGemma Config
* PaliGemma - add separator ind to model/labels
* Refactor + docstring paligemma processor method
* Style
* return toke... | [
{
"path": "src/transformers/models/paligemma/modeling_paligemma.py",
"patch": "@@ -283,9 +283,14 @@ def resize_token_embeddings(self, new_num_tokens: Optional[int] = None, pad_to_m\n self.vocab_size = model_embeds.num_embeddings\n return model_embeds\n \n- def _merge_input_ids_with_image_... | 2024-05-22T17:37:15 |
ollama/ollama | d9bfb2f08fcb18cedc5a036216b770e24fd44006 | 598d6d5572337afac759abf1995950866d96d044 | install: pin fedora to max 37
repos for fedora 38 and newer do not exist as of this commit
```
$ dnf config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/fedora38/x86_64/cuda-fedora38.repo
Adding repo from: https://developer.download.nvidia.com/compute/cuda/repos/fedora38/x86_64/cuda-fed... | [
{
"path": "scripts/install.sh",
"patch": "@@ -231,8 +231,8 @@ if ! check_gpu nvidia-smi || [ -z \"$(nvidia-smi | grep -o \"CUDA Version: [0-9]*\\\n case $OS_NAME in\n centos|rhel) install_cuda_driver_yum 'rhel' $(echo $OS_VERSION | cut -d '.' -f 1) ;;\n rocky) install_cuda_driver_yum 'rh... | 2024-01-16T19:45:12 |
nodejs/node | e509db8c564a660231cc09afef3d900f58201afc | bb59d2bd19cd1e71cc3eeaa4918b36d0980b1c35 | doc: howto decode buffers extending from Writable
Improved stream documentation with an example of how to decode buffers
to strings within a custom Writable.
Fixes: https://github.com/nodejs/node/issues/15369
PR-URL: https://github.com/nodejs/node/pull/16403
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Revi... | [
{
"path": "doc/api/stream.md",
"patch": "@@ -1510,6 +1510,47 @@ class MyWritable extends Writable {\n }\n ```\n \n+#### Decoding buffers in a Writable Stream\n+\n+Decoding buffers is a common task, for instance, when using transformers whose\n+input is a string. This is not a trivial process when using mult... | 2017-10-23T07:28:52 |
rust-lang/rust | 0c4eaa5c985f69d3b6255a56799ee30d29e08d54 | aae7a3c4c2fbd0db49fac5618590df541943f140 | Use default field values for `ErrorOutputType`
Remove manual `Default` impl from `config::ErrorOutputType`. | [
{
"path": "compiler/rustc_session/src/config.rs",
"patch": "@@ -655,10 +655,14 @@ impl OutputType {\n }\n \n /// The type of diagnostics output to generate.\n-#[derive(Clone, Copy, Debug, PartialEq, Eq)]\n+#[derive(Clone, Copy, Debug, PartialEq, Eq, Default)]\n pub enum ErrorOutputType {\n /// Output me... | 2025-03-03T20:18:56 |
golang/go | 9449a125e8b25d21c0d522435be44a4a6e7af2d3 | 3036b76df0ae748856e3e0008b67241cc580e263 | cmd/compile/internal/gc: fix wrong function name in the comment
Change-Id: I2fc5cff7495b5db4eb8f286a5335787241f1a850
GitHub-Last-Rev: 1d226f14c8edbcb189fd1a37b11a82b9642f19ab
GitHub-Pull-Request: golang/go#41917
Reviewed-on: https://go-review.googlesource.com/c/go/+/261317
Reviewed-by: Ian Lance Taylor <iant@golang.or... | [
{
"path": "src/cmd/compile/internal/gc/export.go",
"patch": "@@ -96,7 +96,7 @@ func importsym(ipkg *types.Pkg, s *types.Sym, op Op) *Node {\n \treturn n\n }\n \n-// pkgtype returns the named type declared by symbol s.\n+// importtype returns the named type declared by symbol s.\n // If no such type has been... | 2020-10-11T03:44:09 |
ggml-org/llama.cpp | cc2aa81513ff8bbf8f0527ce36b9366971d5fd35 | 0e219914723695e5e0a1d470c6dbe44e304a8270 | Fix wrong memcpy length for block_interleave == 4 (#19575) | [
{
"path": "ggml/src/ggml-cpu/repack.cpp",
"patch": "@@ -1916,9 +1916,10 @@ static block_q4_Kx8 make_block_q4_Kx8(block_q4_K * in, unsigned int blck_size_in\n int src_offset = (i / 8) * blck_size_interleave;\n int dst_offset = i * blck_size_interleave;\n \n+ // buffer large enough for ... | 2026-02-13T12:32:14 |
huggingface/transformers | d44e1ae03646617b550ff628a706ffd8fe093767 | 0948c827de87056696032f1b84bad309259b0eb7 | Fix link in Pipeline documentation (#30948)
fix documentation as suggested by stevhliu
Co-authored-by: Jun <jun@reliant.ai> | [
{
"path": "src/transformers/pipelines/conversational.py",
"patch": "@@ -270,7 +270,7 @@ def __call__(self, conversations: Union[List[Dict], Conversation, List[Conversat\n Whether or not to clean up the potential extra spaces in the text output.\n generate_kwargs:\n ... | 2024-05-22T16:39:46 |
ollama/ollama | eef50accb425a1815625f496bc5b24d5211d5610 | 05d53de7a10fd95fc475c27b0235b8b51601c2ba | Fix show parameters (#2017) | [
{
"path": "server/routes.go",
"patch": "@@ -15,7 +15,6 @@ import (\n \t\"path/filepath\"\n \t\"reflect\"\n \t\"runtime\"\n-\t\"strconv\"\n \t\"strings\"\n \t\"sync\"\n \t\"syscall\"\n@@ -668,27 +667,12 @@ func GetModelInfo(req api.ShowRequest) (*api.ShowResponse, error) {\n \tcs := 30\n \tfor k, v := range ... | 2024-01-16T18:34:44 |
nodejs/node | bb59d2bd19cd1e71cc3eeaa4918b36d0980b1c35 | 8025bba1a38f7d9b8ba152e45c6a9198f5e49186 | repl: avoid crashing from null and undefined errors
When `throw undefined` or `throw null` is executed, the REPL crashes.
This change does a check for `null|undefined` before accessing an
error's properties to prevent crashing.
Fixes: https://github.com/nodejs/node/issues/16545
Fixes: https://github.com/nodejs/node/i... | [
{
"path": "lib/repl.js",
"patch": "@@ -257,12 +257,13 @@ function REPLServer(prompt,\n }\n } catch (e) {\n err = e;\n- if (err.message === 'Script execution interrupted.') {\n+\n+ if (err && err.message === 'Script execution interrupted.') {\n // The stack trace... | 2017-10-28T16:19:53 |
golang/go | 712cba3bf258c0c75bf1a638cb2119dbb11ed6c7 | d4a5797b8863185230d0e89da9a00fd17f04152a | cmd/go: ignore retracted versions when converting revisions to versions
When a module author retracts a version, the go command should act as
if it doesn't exist unless it's specifically requested.
When converting a revision to a version, we should ignore tags for
retracted versions. For example, if the tag v1.0.0 is... | [
{
"path": "src/cmd/go/internal/modfetch/codehost/codehost.go",
"patch": "@@ -79,9 +79,8 @@ type Repo interface {\n \tReadZip(rev, subdir string, maxSize int64) (zip io.ReadCloser, err error)\n \n \t// RecentTag returns the most recent tag on rev or one of its predecessors\n-\t// with the given prefix and ma... | 2020-10-09T19:16:13 |
ggml-org/llama.cpp | 0e219914723695e5e0a1d470c6dbe44e304a8270 | b2ecc0cdb4cf724b78c770cbbacb8029168d26fb | fix vulkan ggml_acc only works in 3d but not 4d (#19426)
* fix vulkan ggml_acc only works in 3d but not 4d
* removed clamp in test_acc_block
* use the correct stride and its test case
* cuda : fix "supports op" condition
* change src0 to src1 in ggml_vk_acc. Update acc.comp with jeffbolznv\'s suggestion except to ... | [
{
"path": "ggml/src/ggml-cuda/ggml-cuda.cu",
"patch": "@@ -4822,8 +4822,11 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g\n case GGML_OP_CONV_2D_DW:\n case GGML_OP_CONV_TRANSPOSE_2D:\n case GGML_OP_POOL_2D:\n- case GGML_OP_ACC:\n ... | 2026-02-13T12:31:37 |
huggingface/transformers | b1065aa08ac0da11fcb9e3827cd7eafabe4beebd | 1d568dfab262f76079eb4f3d05b606d51a0c9e4b | Generation: get special tokens from model config (#30899)
* fix
* let's do this way?
* codestyle
* update
* add tests | [
{
"path": "src/transformers/generation/utils.py",
"patch": "@@ -1361,6 +1361,23 @@ def _get_cache(self, cache_implementation: str, max_batch_size: int, max_cache_l\n self._cache.reset()\n return self._cache\n \n+ def _get_decoder_start_token_id(\n+ self, decoder_start_token_id:... | 2024-05-22T16:15:41 |
nodejs/node | dfcaf280b728b20e536c27964ae1446342d98607 | c396fbfedf62cdd77a0261835d3ec448911a187f | src: fix etw provider include on Windows
PR-URL: https://github.com/nodejs/node/pull/16639
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com> | [
{
"path": "src/node_win32_etw_provider.cc",
"patch": "@@ -19,6 +19,7 @@\n // OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE\n // USE OR OTHER DEALINGS IN THE SOFTWARE.\n \n+#include \"node_win32_etw_provider.h\"\n #include \"node_etw_provider.h\"\n #include \"node_win32_etw_provid... | 2017-10-31T12:27:18 |
golang/go | 33511fb959f8f0edd5e831a4b41523daf9d84e87 | 2be7788f8383c2330cd96db53273e2995d4468f8 | net/http/pprof: remove html/template dependency
html/template indirectly uses reflect MethodByName, this causes linker
to use conservative mode resulting in larger binaries. The template here
is trivial and can be replaced by string manipulation.
This reduces a binary using only net/http/pprof by ~2.5MB.
Fixes #4156... | [
{
"path": "src/go/build/deps_test.go",
"patch": "@@ -449,7 +449,7 @@ var depsRules = `\n \tOS, compress/gzip, regexp\n \t< internal/profile;\n \n-\thtml/template, internal/profile, net/http, runtime/pprof, runtime/trace\n+\thtml, internal/profile, net/http, runtime/pprof, runtime/trace\n \t< net/http/pprof;... | 2020-09-23T11:06:28 |
Subsets and Splits
Assembly Language GitHub Issues
Retrieves a sample of assembly-language related commits with their details, but doesn't provide meaningful analysis or patterns beyond basic filtering.
Swift Compiler Issues Analysis
Retrieves all training data for the Swift programming language repository, providing basic filtering but offering limited analytical insight beyond identifying relevant code examples.