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 |
|---|---|---|---|---|---|
huggingface/transformers | 9839d57a0244f86125c89981f6301b48309b4913 | e85d5ab2bb7e45f8fbcbef5ff1e66c2997de855c | Fix serving continuous batching (#41624)
* udpate-serving-cb
* style
* style
* check none
* Apply suggestions from code review
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com> | [
{
"path": "docs/source/en/serving.md",
"patch": "@@ -380,7 +380,7 @@ CB is opt-in and currently applies to chat completions.\n ```sh\n transformers serve \\\n --continuous-batching\n- --attn_implementation sdpa_paged\n+ --attn_implementation \"sdpa\"\n ```\n \n ### Performance tips\n@@ -390,11 +390,10 @... | 2025-10-16T15:24:21 |
ollama/ollama | 92189a58553ac88edf19ea5a8834478bd5fd6d86 | d790bf99166d5a97d9d01c78b0658706c28580b5 | fix memory check | [
{
"path": "api/client.go",
"patch": "@@ -14,6 +14,7 @@ import (\n \t\"runtime\"\n \t\"strings\"\n \n+\t\"github.com/jmorganca/ollama/format\"\n \t\"github.com/jmorganca/ollama/version\"\n )\n \n@@ -127,7 +128,7 @@ func (c *Client) do(ctx context.Context, method, path string, reqData, respData\n \treturn nil... | 2023-10-12T16:34:16 |
vuejs/vue | 4cf49828c0274262f111e23056332d3a970b28d9 | 2a5fb41d1cd7445b6aacd2e888df1b397b1899dd | properly handle cosntructor options modification before global mixin application (fix #4976) | [
{
"path": "flow/component.js",
"patch": "@@ -14,6 +14,7 @@ declare interface Component {\n static extend: (options: Object) => Function;\n static superOptions: Object;\n static extendOptions: Object;\n+ static sealedOptions: Object;\n static super: Class<Component>;\n // assets\n static directi... | 2017-02-23T23:11:18 |
huggingface/transformers | e85d5ab2bb7e45f8fbcbef5ff1e66c2997de855c | 1c36d407d5ce0712c702e4314c62705c012bb700 | Fix dtype casting with quantization (#41665)
fix dtype casting | [
{
"path": "src/transformers/modeling_utils.py",
"patch": "@@ -557,7 +557,7 @@ def _infer_parameter_dtype(\n is_param_float8_e4m3fn = is_torch_e4m3fn_available and empty_param.dtype == torch.float8_e4m3fn\n if empty_param.dtype.is_floating_point and not is_param_float8_e4m3fn:\n # dtype that ... | 2025-10-16T15:19:32 |
ggml-org/llama.cpp | 120bf7046d85a893f064c12abe58bfeebd735f84 | 4258e0cfe72c72ad2787be1e9f93253e89219455 | CUDA + openCL: fix bug in accessing rms_norm->src while doing fusion (#16577) | [
{
"path": "ggml/src/ggml-cuda/ggml-cuda.cu",
"patch": "@@ -2876,7 +2876,7 @@ static bool ggml_cuda_can_fuse(const struct ggml_cgraph * cgraph, int node_idx,\n }\n \n //if rms norm is the B operand, then we don't handle broadcast\n- if (rms_norm == mul->src[1] && !ggml_are_same_shape(m... | 2025-10-14T14:48:08 |
denoland/deno | 812d567c31638cd5a64ced5909be2ba25f08d50e | b715bb8ec9016c009e74d3982ab0b843b961839a | fix: use `locked-tripwire` to prevent unlocked cargo installs (#31973) | [
{
"path": "Cargo.lock",
"patch": "@@ -1659,6 +1659,7 @@ dependencies = [\n \"libc\",\n \"libsui\",\n \"libz-sys\",\n+ \"locked-tripwire\",\n \"log\",\n \"lol_html\",\n \"lsp-types\",\n@@ -5967,6 +5968,12 @@ dependencies = [\n \"scopeguard\",\n ]\n \n+[[package]]\n+name = \"locked-tripwire\"\n+version... | 2026-01-29T02:26:35 |
vuejs/vue | 2a5fb41d1cd7445b6aacd2e888df1b397b1899dd | 01b09e6bcdbfadd45c014a6be4f9fca5e3896969 | re-implement mergeVNodeHook to prevent memory leak (fix #4990) | [
{
"path": "src/core/vdom/helpers/merge-hook.js",
"patch": "@@ -1,18 +1,34 @@\n /* @flow */\n \n-export function mergeVNodeHook (def: Object, hookKey: string, hook: Function, key: string) {\n- key = key + hookKey\n- const injectedHash: Object = def.__injected || (def.__injected = {})\n- if (!injectedHash[... | 2017-02-23T00:02:47 |
ggml-org/llama.cpp | 1ee9d0b415cdf5240418c110a18b419f4002b154 | 48e2fa9fb7c2de1e53808fdb65ec33f916020fc4 | CUDA: use fastdiv + ggml_cuda_mad for mmvf (#16557)
* CUDA: use fastdiv + ggml_cuda_mad for mmvf
* use bf16 directly + fix formatting
* Add exception for HIP code | [
{
"path": "ggml/src/ggml-cuda/mmvf.cu",
"patch": "@@ -7,14 +7,14 @@ template <typename T, typename type_acc, int ncols_dst, int block_size>\n static __global__ void mul_mat_vec_f(\n const T * __restrict__ x, const float * __restrict__ y, const int32_t * __restrict__ ids, float * __restrict__ dst,\n ... | 2025-10-14T11:16:21 |
denoland/deno | b715bb8ec9016c009e74d3982ab0b843b961839a | 591e9267e6e052d6d7b11a7a5be233b73a27f121 | fix: allow reading /dev/tty without requiring --allow-all (#31105) | [
{
"path": "runtime/permissions/lib.rs",
"patch": "@@ -4025,7 +4025,7 @@ impl PermissionsContainer {\n })\n } else {\n let path = self.descriptor_parser.parse_special_file_descriptor(path)?;\n- self.check_special_file(path, api_name)\n+ self.check_special_file(path, access_kind, api... | 2026-01-28T17:21:25 |
ollama/ollama | 6fe178134dd27c6aaf64374f4163abf6c8d624dd | d890890f665b5220ec1f35829cb7352abdb19c0a | improve api error handling (#781)
- remove new lines from llama.cpp error messages relayed to client
- check api option types and return error on wrong type
- change num layers from 95% VRAM to 92% VRAM | [
{
"path": "api/types.go",
"patch": "@@ -3,7 +3,6 @@ package api\n import (\n \t\"encoding/json\"\n \t\"fmt\"\n-\t\"log\"\n \t\"math\"\n \t\"os\"\n \t\"reflect\"\n@@ -238,44 +237,39 @@ func (opts *Options) FromMap(m map[string]interface{}) error {\n \t\t\t\t\t\t// when JSON unmarshals numbers, it uses float6... | 2023-10-13T20:57:10 |
vuejs/vue | 01b09e6bcdbfadd45c014a6be4f9fca5e3896969 | 73c5614740bee7c439735a7cf57841981a2b9ac2 | type for renderError option | [
{
"path": "types/options.d.ts",
"patch": "@@ -22,6 +22,7 @@ export interface ComponentOptions<V extends Vue> {\n el?: Element | String;\n template?: string;\n render?(this: V, createElement: CreateElement): VNode;\n+ renderError?: (h: () => VNode, err: Error) => VNode;\n staticRenderFns?: ((createE... | 2017-02-22T18:17:18 |
huggingface/transformers | 1c36d407d5ce0712c702e4314c62705c012bb700 | 0215846d98e25a39952530cbdf7cc42bd4515324 | Add in-out modalities as class attribute per model (#41366)
* update all models
* fix copies
* explanation comment
* better notation in omni model
* style
* fix copies
* output_modalities under generation mixin
* fix copies
* oh, glm4v also needs conversion | [
{
"path": "src/transformers/generation/utils.py",
"patch": "@@ -366,6 +366,9 @@ class GenerationMixin(ContinuousMixin):\n To learn more about decoding strategies refer to the [text generation strategies guide](../generation_strategies).\n \"\"\"\n \n+ # Should be overwritten by models that can ge... | 2025-10-16T15:11:06 |
denoland/deno | 591e9267e6e052d6d7b11a7a5be233b73a27f121 | 02c2406524b26f10a70947fcb4ac022fa1bbe2fd | fix(ext/node): use primordials in `_fs_ftruncate.ts` (#31944)
Towards #24236
<!--
Before submitting a PR, please read
https://docs.deno.com/runtime/manual/references/contributing
1. Give the PR a descriptive title.
Examples of good title:
- fix(std/http): Fix race condition in server
- docs(console): Upda... | [
{
"path": "ext/node/polyfills/_fs/_fs_ftruncate.ts",
"patch": "@@ -1,10 +1,15 @@\n // Copyright 2018-2026 the Deno authors. MIT license.\n \n-// TODO(petamoriken): enable prefer-primordials for node polyfills\n-// deno-lint-ignore-file prefer-primordials\n-\n-import { CallbackWithError } from \"ext:deno_nod... | 2026-01-28T14:45:05 |
ggml-org/llama.cpp | 48e2fa9fb7c2de1e53808fdb65ec33f916020fc4 | 5b6913c47b6bc71a6f927805a45387d5657d8b89 | CUDA: add fp kernel for larger batch size MoE (#16512)
* CUDA: kernel for larger batch sizes for MoE
* WIP
* WIP
* WIP
* WIP
* WIP
* WIP
* fixup
* tests
* Move mmq_ids_helper to mmid
* cleanup
* Remove redundant checks | [
{
"path": "ggml/src/ggml-cuda/mmf.cu",
"patch": "@@ -1,5 +1,7 @@\n #include \"ggml.h\"\n #include \"mmf.cuh\"\n+#include \"mmid.cuh\"\n+\n \n void ggml_cuda_mul_mat_f(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * ids, ggml_tensor * dst) {\n GGML... | 2025-10-14T11:15:15 |
vuejs/vue | 8e854a9ed1b606890b53637f201432174bb7508a | 29f6902d1959c35eb38ef110bee6c13124903442 | only use click for v-model checkbox/radio in Chrome (fix #4796, #4896) | [
{
"path": "src/core/util/env.js",
"patch": "@@ -14,6 +14,7 @@ export const isIE9 = UA && UA.indexOf('msie 9.0') > 0\n export const isEdge = UA && UA.indexOf('edge/') > 0\n export const isAndroid = UA && UA.indexOf('android') > 0\n export const isIOS = UA && /iphone|ipad|ipod|ios/.test(UA)\n+export const isC... | 2017-02-22T02:25:57 |
huggingface/transformers | bf815e9b5ea076f758cc58f73f2be2d36237f9ec | 4a43e3d57cde839b7226750879b30d34a5de4598 | [`Masks`] Fix mask handling in eager for vision models (#41625)
add mask handling in case of models that do use it | [
{
"path": "src/transformers/models/audio_spectrogram_transformer/modeling_audio_spectrogram_transformer.py",
"patch": "@@ -96,25 +96,28 @@ def forward(self, input_values: torch.Tensor) -> torch.Tensor:\n return embeddings\n \n \n-# Copied from transformers.models.vit.modeling_vit.eager_attention_for... | 2025-10-16T14:27:26 |
ollama/ollama | 257ffeb99747a2bb9caed54b1d51020aca95b0ae | 92578798bb1abcedd6bc99479d804f32d9ee2f6c | fix download | [
{
"path": "server/download.go",
"patch": "@@ -144,7 +144,7 @@ func (b *blobDownload) run(ctx context.Context, requestURL *url.URL, opts *Regis\n \n \tfile.Truncate(b.Total)\n \n-\tg, _ := errgroup.WithContext(ctx)\n+\tg, inner := errgroup.WithContext(ctx)\n \tg.SetLimit(numDownloadParts)\n \tfor i := range ... | 2023-10-12T19:52:35 |
denoland/deno | 02c2406524b26f10a70947fcb4ac022fa1bbe2fd | f27a7984cdf23d458b05f2ed54cd6b20e173baae | fix(ext/node): fix usage of new V8 string APIs (#31963)
Co-authored-by: Divy Srivastava <me@littledivy.com> | [
{
"path": "ext/napi/js_native_api.rs",
"patch": "@@ -2129,16 +2129,12 @@ fn napi_get_value_string_latin1(\n *result = value.length();\n }\n } else if bufsize != 0 {\n- let buffer =\n- unsafe { std::slice::from_raw_parts_mut(buf as _, bufsize - 1) };\n- value.write_one_byte_v2(scope, 0... | 2026-01-28T12:26:45 |
ggml-org/llama.cpp | e38b7c6e9e4453e3b3e96d76e38bc2ccb6bce458 | 5016b7286240d29f8f640039989b84ea3a854344 | graph : support cacheless embeddings with FA and iSWA (#16528)
* graph : support cacheless embeddings with FA and iSWA
* cont : deduplicate mask creation
* cont : fix name | [
{
"path": "src/llama-graph.cpp",
"patch": "@@ -261,12 +261,17 @@ void llm_graph_input_cross_embd::set_input(const llama_ubatch * ubatch) {\n }\n }\n \n-static void print_mask(float * data, int64_t n_tokens, int64_t n_kv, int64_t n_swa, llama_swa_type swa_type) {\n+static void print_mask(const float * da... | 2025-10-13T19:42:37 |
vuejs/vue | 1861ee9570730149e01f225323c3a52392e5900f | a1d5b22e347aab9e8e4f63b01abf2d88dec428df | feat: renderError | [
{
"path": "flow/options.js",
"patch": "@@ -33,7 +33,8 @@ declare type ComponentOptions = {\n // DOM\n el?: string | Element;\n template?: string;\n- render: () => VNode;\n+ render: (h: () => VNode) => VNode;\n+ renderError?: (h: () => VNode, err: Error) => VNode;\n staticRenderFns?: Array<() => V... | 2017-02-20T05:16:40 |
huggingface/transformers | 8725ce10edb29771fb9a1aa108e6a04859efe973 | 1fb3fc4db0e87fd7c2f57a36b6b32ee6fa69c50c | [Fix] Deepseek V3 expert bias routing (#41647)
* [Fix] Deepseek V3 expert bias routing
* [Fix] fix-copies
* [Fix] Run make style | [
{
"path": "src/transformers/models/deepseek_v3/modeling_deepseek_v3.py",
"patch": "@@ -176,9 +176,11 @@ def __init__(self, config):\n \n def route_tokens_to_experts(self, router_logits):\n router_logits = router_logits.sigmoid()\n- router_logits = router_logits + self.gate.e_score_correct... | 2025-10-16T14:04:48 |
ollama/ollama | 56497663c8bc7de73f4ff8d53235664a044c9d95 | 92578798bb1abcedd6bc99479d804f32d9ee2f6c | relay model runner error message to client (#720)
* give direction to user when runner fails
* also relay errors from timeout
* increase timeout to 3 minutes | [
{
"path": "cmd/cmd.go",
"patch": "@@ -479,18 +479,7 @@ func generate(cmd *cobra.Command, model, prompt string, wordWrap bool) error {\n \t}\n \n \tif err := client.Generate(cancelCtx, &request, fn); err != nil {\n-\t\tif strings.Contains(err.Error(), \"failed to load model\") {\n-\t\t\t// tell the user to c... | 2023-10-12T15:16:37 |
ggml-org/llama.cpp | 5016b7286240d29f8f640039989b84ea3a854344 | 7049736b2dd9011bf819e298b844ebbc4b5afdc9 | opencl: fix build targeting CL 2 (#16554) | [
{
"path": "ggml/src/ggml-opencl/ggml-opencl.cpp",
"patch": "@@ -2348,8 +2348,13 @@ static ggml_backend_opencl_context * ggml_cl2_init(ggml_backend_dev_t dev) {\n svm_caps & CL_DEVICE_SVM_ATOMICS ? \"true\" : \"false\");\n \n if (opencl_c_version.major >= 3) {\n+ // Assume it is not availa... | 2025-10-13T18:50:37 |
denoland/deno | 0864b5442d25d7ac8580c8b9b9adb3629c07667a | a1fb6b6d0a3e9910f9a9915ee300b5be429f6cb7 | fix(ext/node): read NODE_OPTIONS for --require and --inspect-publish-uid (#31949) | [
{
"path": "Cargo.lock",
"patch": "@@ -1667,6 +1667,7 @@ dependencies = [\n \"memchr\",\n \"nix 0.27.1\",\n \"node_resolver\",\n+ \"node_shim\",\n \"notify\",\n \"once_cell\",\n \"open\",",
"additions": 1,
"deletions": 0,
"language": "Unknown"
},
{
"path": "cli/Cargo.toml",
"pat... | 2026-01-28T03:10:28 |
vuejs/vue | a1d5b22e347aab9e8e4f63b01abf2d88dec428df | 3c0cdb553590db84a3bdb2a299947e7424e23464 | throw error in ssr so it can be captured properly by renderer | [
{
"path": "src/core/util/error.js",
"patch": "@@ -1,5 +1,6 @@\n import config from '../config'\n import { warn } from './debug'\n+import { inBrowser } from './env'\n \n export function handleError (err, vm, type) {\n if (config.errorHandler) {\n@@ -8,8 +9,10 @@ export function handleError (err, vm, type) ... | 2017-02-20T04:21:02 |
huggingface/transformers | 503c933f360fd3e577d9db21fed02253ecfdf7d5 | 2aff20aff6119362e56aede7bf6846d1b037a41f | Fix confusing cls assignment (#41642)
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com> | [
{
"path": "src/transformers/utils/fx.py",
"patch": "@@ -777,9 +777,9 @@ def __new__(\n attrs: dict[str, Any],\n proxy_factory_fn: Optional[Callable[[Node], Proxy]] = None,\n ):\n- cls = super().__new__(cls, name, bases, attrs)\n- for attr_name in dir(cls):\n- att... | 2025-10-16T13:01:07 |
ggml-org/llama.cpp | 7049736b2dd9011bf819e298b844ebbc4b5afdc9 | 01d2bdc2bc61b676706830305b286b08b9885a41 | CUDA: fix numerical issues in tile FA kernel (#16540) | [
{
"path": "ggml/src/ggml-cuda/fattn-tile.cuh",
"patch": "@@ -540,10 +540,12 @@ static __device__ __forceinline__ void flash_attn_tile_iter(\n KQ_acc[(i_KQ_0/(np*warp_size))*cpw + jc0] = logit_softcap * tanhf(KQ_acc[(i_KQ_0/(np*warp_size))*cpw + jc0]);\n }\n \n- KQ_acc[... | 2025-10-13T14:29:45 |
ollama/ollama | 92578798bb1abcedd6bc99479d804f32d9ee2f6c | 788637918ad76d2639114c5736e9d70b94758c4e | fix relative links in `README.md` | [
{
"path": "README.md",
"patch": "@@ -13,7 +13,7 @@ Get up and running with large language models locally.\n \n ### macOS\n \n-[Download](https://ollama.ai/download/Ollama-darwin.zip) \n+[Download](https://ollama.ai/download/Ollama-darwin.zip)\n \n ### Linux & WSL2\n \n@@ -23,7 +23,7 @@ curl https://ollama.a... | 2023-10-11T23:24:06 |
denoland/deno | a1fb6b6d0a3e9910f9a9915ee300b5be429f6cb7 | 7768451f225482d05b072289feffa12d5ba9a978 | fix(ext/node): pass test-child-process-bad-stdio node compat test (#31851)
Added shell command transformation logic, which handles executable
path detection (including shell variable references like "${VAR}"),
argument splitting, and translation.
Enable test-child-process-bad-stdio.js node compat test. | [
{
"path": "ext/node/polyfills/internal/child_process.ts",
"patch": "@@ -970,7 +970,7 @@ export function normalizeSpawnArguments(\n if (options.shell) {\n let command = ArrayPrototypeJoin([file, ...args], \" \");\n // Transform Node.js flags to Deno equivalents in shell commands that invoke Deno\n-... | 2026-01-28T03:07:38 |
vuejs/vue | 3c0cdb553590db84a3bdb2a299947e7424e23464 | 3566d9296de28e679be0970c4eb52a938eaf2ecc | improve error handling for lifecycle hooks | [
{
"path": "src/core/instance/lifecycle.js",
"patch": "@@ -6,7 +6,7 @@ import { createEmptyVNode } from '../vdom/vnode'\n import { observerState } from '../observer/index'\n import { updateComponentListeners } from './events'\n import { resolveSlots } from './render-helpers/resolve-slots'\n-import { warn, va... | 2017-02-20T04:19:44 |
huggingface/transformers | 2aff20aff6119362e56aede7bf6846d1b037a41f | 981370c038e98f4a4c720aa6b56e5cbde8266c74 | Fix typos in documentation (#41641)
Fix typos
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com> | [
{
"path": "docs/source/en/model_doc/evolla.md",
"patch": "@@ -61,7 +61,7 @@ message_list = [\n ]\n ]\n input_dict = processor(\n- protein_informations, messages_list, return_tensors=\"pt\", text_max_length=512, protein_max_length=1024\n+ protein_inputs, messages_list, return_tensors=\"pt\", text_m... | 2025-10-16T12:58:46 |
ggml-org/llama.cpp | 01d2bdc2bc61b676706830305b286b08b9885a41 | 56fc38b9655fbe1869d8bd6cfb269418196cea69 | ggml : fix build broken with -march=armv9-a on MacOS (#16520)
* ggml : fix build broken with -march=armv9-a on MacOS
Signed-off-by: Jie Fu <jiefu@tencent.com>
* Add #pragma message
Signed-off-by: Jie Fu <jiefu@tencent.com>
* Address review comment.
Signed-off-by: Jie Fu <jiefu@tencent.com>
* Update ggml/src/ggml... | [
{
"path": "ggml/src/ggml-cpu/ggml-cpu-impl.h",
"patch": "@@ -68,7 +68,7 @@ struct ggml_compute_params {\n #endif // __VXE2__\n #endif // __s390x__ && __VEC__\n \n-#if defined(__ARM_FEATURE_SVE)\n+#if defined(__ARM_FEATURE_SVE) && defined(__linux__)\n #include <sys/prctl.h>\n #endif\n ",
"additions": 1... | 2025-10-13T12:48:47 |
denoland/deno | 824b30c99682377b567b12a73b5ef21d376c17af | 3ae0124062ab3340ef13d69e36d5644d7f32d65d | fix: graceful server shutdown with open, un-upgraded connections (#31959) | [
{
"path": "ext/http/http_next.rs",
"patch": "@@ -874,7 +874,8 @@ async fn serve_http2_autodetect(\n options: Options,\n ) -> Result<(), HttpNextError> {\n let prefix = NetworkStreamPrefixCheck::new(io, HTTP2_PREFIX);\n- let (matches, io) = prefix.match_prefix().await?;\n+ let (matches, io) =\n+ pre... | 2026-01-27T15:30:27 |
vuejs/vue | 3dc9338479c6c04925831b21ebec43677a918897 | 9ccffe72db6d4c35a137063519e6ec0a9c5c34a4 | Warn when defining a method with same name as a prop (#4950)
* Warn when defining a method with same name as a prop
* update error message | [
{
"path": "src/core/instance/state.js",
"patch": "@@ -183,14 +183,23 @@ function createComputedGetter (key) {\n }\n \n function initMethods (vm: Component, methods: Object) {\n+ const props = vm.$options.props\n for (const key in methods) {\n vm[key] = methods[key] == null ? noop : bind(methods[key],... | 2017-02-18T14:09:27 |
huggingface/transformers | 981370c038e98f4a4c720aa6b56e5cbde8266c74 | eef9fb2af3db888cf93f81b425f9db453336726c | Format MarkDown documentation and tiny fixes (#41638)
* Fix MarkDown syntax
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com>
* More fixes
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com>
---------
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com> | [
{
"path": "docs/source/en/accelerator_selection.md",
"patch": "@@ -55,6 +55,7 @@ deepspeed --num_gpus 2 trainer-program.py ...\n </hfoptions>\n \n ## Order of accelerators\n+\n To select specific accelerators to use and their order, use the environment variable appropriate for your hardware. This is often s... | 2025-10-16T12:58:06 |
ggml-org/llama.cpp | 56fc38b9655fbe1869d8bd6cfb269418196cea69 | 1fb9504eb744969a990bfe4cfcf1d3d7a479541c | CANN: fix CPU memory leak in CANN backend (#16549)
This commit fixes a CPU-side memory leak issue in the CANN backend,
which occurred when intermediate aclTensorList objects were not properly
released after operator execution. The leak happened during repeated
invocations of CANN ops (e.g., FlashAttention), leading to... | [
{
"path": "ggml/src/ggml-cann/aclnn_ops.cpp",
"patch": "@@ -146,9 +146,7 @@ void ggml_cann_op_unary_gated(\n unary_op(ctx, acl_src0, acl_dst);\n GGML_CANN_CALL_ACLNN_OP(ctx, InplaceMul, acl_dst, acl_src1);\n \n- ggml_cann_release_resources(ctx, acl_src0, acl_dst);\n- if(src1)\n- ggml_ca... | 2025-10-13T09:01:24 |
vuejs/vue | 2f520ba9eeee24479536b9621cf9b32cd463cc35 | 7f6c889a8811b29e571bba7b39f21284c090fa94 | fix doc comment typo (#4954) | [
{
"path": "src/core/instance/state.js",
"patch": "@@ -58,7 +58,7 @@ function initProps (vm: Component, propsOptions: Object) {\n const propsData = vm.$options.propsData || {}\n const props = vm._props = {}\n // cache prop keys so that future props updates can iterate using Array\n- // instead of dyan... | 2017-02-17T09:31:45 |
denoland/deno | 3ae0124062ab3340ef13d69e36d5644d7f32d65d | 1b5ba4c0e0e188351d470883115f5ddca0ea47f8 | fix(unstable): lint ast comments from first file showing in others (#31956)
This issue was caused by capturing the first ever invoked `ctx` variable
instead of accessing the `ctx` of the file that is currently being
linted.
Fixes #31896 | [
{
"path": "cli/js/40_lint.js",
"patch": "@@ -631,8 +631,8 @@ function setNodeGetters(ctx) {\n hasCommenstGetter = true;\n Object.defineProperty(FacadeNode.prototype, \"comments\", {\n get() {\n- materializeComments(ctx);\n- return ctx.comments;\n+ materializeComments(this[... | 2026-01-27T13:17:31 |
huggingface/transformers | eef9fb2af3db888cf93f81b425f9db453336726c | 35dc8f0a2e07239633e5d44147af5014a67e43c4 | Fix EncoderDecoder cache (#41612)
* Fix EncoderDecoder cache
* Add the option for the ddp data tuples to have 2 elems
* Modifiy the order of the KV and sliding
* Adapted RAG and Whisper to new EncoderDecoderCache
* A single comma
* Remove kwargs in map
* Fixed order in manual injection cache test
* Slight chang... | [
{
"path": "src/transformers/cache_utils.py",
"patch": "@@ -937,7 +937,7 @@ class DynamicCache(Cache):\n \n def __init__(\n self,\n- ddp_cache_data: Optional[Iterable[tuple[Optional[torch.Tensor], torch.Tensor, torch.Tensor]]] = None,\n+ ddp_cache_data: Optional[Iterable[tuple[Optio... | 2025-10-16T12:55:41 |
ggml-org/llama.cpp | 1fb9504eb744969a990bfe4cfcf1d3d7a479541c | 3f750f8d760ab5a61491e6a9409072dfeee4b4d7 | fix: add remark plugin to render raw HTML as literal text (#16505)
* fix: add remark plugin to render raw HTML as literal text
Implemented a missing MDAST stage to neutralize raw HTML like major LLM WebUIs
do ensuring consistent and safe Markdown rendering
Introduced 'remarkLiteralHtml', a plugin that converts raw H... | [
{
"path": "tools/server/webui/package-lock.json",
"patch": "@@ -50,6 +50,7 @@\n \t\t\t\t\"eslint-plugin-svelte\": \"^3.0.0\",\n \t\t\t\t\"fflate\": \"^0.8.2\",\n \t\t\t\t\"globals\": \"^16.0.0\",\n+\t\t\t\t\"mdast\": \"^3.0.0\",\n \t\t\t\t\"mdsvex\": \"^0.12.3\",\n \t\t\t\t\"playwright\": \"^1.53.0\",\n \t\... | 2025-10-13T08:55:32 |
ollama/ollama | c413a5509313b6c9800d60ecf926a368f141d42b | 630bb75d2acd2ae6bc5182e2f53dc8b27b7877b3 | download: handle inner errors | [
{
"path": "server/download.go",
"patch": "@@ -31,11 +31,13 @@ type blobDownload struct {\n \n \tTotal int64\n \tCompleted atomic.Int64\n-\tdone bool\n \n \tParts []*blobDownloadPart\n \n \tcontext.CancelFunc\n+\n+\tdone bool\n+\terr error\n \treferences atomic.Int32\n }\n \n@@ -125,7 +... | 2023-10-11T20:49:01 |
vuejs/vue | 0598ab0c6438cdd07ba7d7bc0cbdbc19231f26f8 | acec8db2c4b54ac6f9cda7ab8fcc3bdd213eb3ed | #4371 - Explicit transition durations (#4857)
* Add transition explicit duration
* Fix tests for explicit transition duration
* Tweaks & default to milliseconds
* Better tests
* Better test for change value case
* Fix transition duration tests
* Better flow typing
* Fix transition test
* Revert... | [
{
"path": "src/platforms/web/runtime/components/transition.js",
"patch": "@@ -21,7 +21,8 @@ export const transitionProps = {\n leaveActiveClass: String,\n appearClass: String,\n appearActiveClass: String,\n- appearToClass: String\n+ appearToClass: String,\n+ duration: [Number, Object]\n }\n \n // i... | 2017-02-15T23:37:09 |
huggingface/transformers | 35dc8f0a2e07239633e5d44147af5014a67e43c4 | 2935a1be19f12176c455cb65d67dc5a3bb84cd77 | Adjust device logging level and add minor fixes (#41636)
This commit addresses a noisy warning and improves the robustness of the base pipeline implementation.
- The device placement message in the pipeline base class has been changed from a `warning` to a `debug` log. This reduces log noise for users who are aware o... | [
{
"path": "src/transformers/pipelines/base.py",
"patch": "@@ -94,6 +94,7 @@ def _pad(items, key, padding_value, padding_side):\n min_length = min(item[key].shape[1] for item in items)\n dtype = items[0][key].dtype\n \n+ tensor = None\n if dim == 2:\n if max_length ... | 2025-10-16T12:47:39 |
denoland/deno | 9bd69275af53dd32bc55422b96bb9efae81fba6b | 6b4468a841a55ffdc5c447a53b9a97bbdf4a2cc1 | fix(ext/net): remove socket file when dropping unix listener (#31947) | [
{
"path": "ext/net/ops_unix.rs",
"patch": "@@ -47,6 +47,34 @@ impl Resource for UnixDatagramResource {\n }\n }\n \n+/// A wrapper around `UnixListener` that holds the socket path and removes it on drop.\n+pub struct UnixListenerWithPath {\n+ listener: UnixListener,\n+ path: PathBuf,\n+}\n+\n+impl UnixLi... | 2026-01-27T02:30:11 |
ggml-org/llama.cpp | c515fc577166042234241c6bd0da9b08dcbe2bb9 | f9bc66c3ebcfddb5f09e4b21253623caeb8e414a | ggml : fix scalar path for computing norm (#16558) | [
{
"path": "ggml/src/ggml-cpu/vec.cpp",
"patch": "@@ -463,9 +463,9 @@ ggml_float ggml_vec_cvar_f32(const int n, float * y, const float * x, const floa\n #endif\n for (; i < n; ++i) {\n float val = x[i] - mean;\n+ y[i] = val;\n val *= val;\n sum += (ggml_float)val;\n- ... | 2025-10-13T08:22:27 |
ollama/ollama | ab0668293cbfc2188b736d4c2b7dc0b7b997f5bf | af4cf55884ac54b9e637cd71dadfe9b7a5685877 | llm: fix build on `amd64` | [
{
"path": "llm/llama.cpp/generate_darwin_amd64.go",
"patch": "@@ -15,4 +15,4 @@ package llm\n //go:generate git -C gguf apply ../patches/0001-remove-warm-up-logging.patch\n //go:generate cmake -S gguf -B gguf/build/cpu -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=x86_64 -DCMAKE_OSX_ARC... | 2023-10-06T21:39:54 |
huggingface/transformers | 2935a1be19f12176c455cb65d67dc5a3bb84cd77 | b9bd8c45a15b4e90bd800dcf6857cd6d690b7ca4 | Fix fp32_ln for various models (#41605)
* Add is_causal to KosmosTextAttention
* Move get target_dtype to be imported elsewhere
* Fix fp32 flash attention bug in bark
* Fix is_causal in mllama
* Fix fp32 issue on StableLM
* Fix repo-consistency | [
{
"path": "src/transformers/integrations/flash_attention.py",
"patch": "@@ -11,6 +11,19 @@\n _use_top_left_mask = flash_attn_supports_top_left_mask()\n \n \n+def get_target_dtype(query: torch.Tensor, module: torch.nn.Module) -> torch.dtype:\n+ \"\"\"If the query is in float32, return a target dtype compa... | 2025-10-16T12:18:49 |
vuejs/vue | acec8db2c4b54ac6f9cda7ab8fcc3bdd213eb3ed | 8d885128371c81994715691c81d161b3768706dd | [WIP] Support for ref callback (#4807)
* :sparkles: Tests for ref callback
* :sparkles: Support for ref callback
* Add test of inline ref callback
* adjust ref implementation strategy
* fix patch ref registration
* fix tests
* fix flow
* fix test for phantomjs | [
{
"path": "flow/component.js",
"patch": "@@ -2,6 +2,10 @@ import type { Config } from '../src/core/config'\n import type VNode from '../src/core/vdom/vnode'\n import type Watcher from '../src/core/observer/watcher'\n \n+declare type Refs = {\n+ [key: string]: Component | Element | Array<Component | Element... | 2017-02-15T22:29:13 |
ggml-org/llama.cpp | 81d54bbfd599811b354c39f04550888168be7780 | c7be9febcbafa9af7d1b9443f86475c59c9c5f87 | webui: remove client-side context pre-check and rely on backend for limits (#16506)
* fix: make SSE client robust to premature [DONE] in agentic proxy chains
* webui: remove client-side context pre-check and rely on backend for limits
Removed the client-side context window pre-check and now simply sends messages
whi... | [
{
"path": "tools/server/webui/src/lib/components/app/chat/ChatScreen/ChatScreen.svelte",
"patch": "@@ -7,6 +7,7 @@\n \t\tChatMessages,\n \t\tChatProcessingInfo,\n \t\tEmptyFileAlertDialog,\n+\t\tChatErrorDialog,\n \t\tServerErrorSplash,\n \t\tServerInfo,\n \t\tServerLoadingSplash,\n@@ -22,10 +23,11 @@\n \t\... | 2025-10-12T16:06:41 |
denoland/deno | 6b4468a841a55ffdc5c447a53b9a97bbdf4a2cc1 | a9b7c7d6d0c452a6697e8b8bbab89f427656c238 | fix(ext/http): use serve address override only once (#31935) | [
{
"path": "ext/http/00_serve.ts",
"patch": "@@ -698,6 +698,9 @@ function formatHostName(hostname: string): string {\n return StringPrototypeIncludes(hostname, \":\") ? `[${hostname}]` : hostname;\n }\n \n+// Flag to track if DENO_SERVE_ADDRESS override has been consumed\n+let serveAddressOverrideConsumed ... | 2026-01-27T00:42:27 |
ollama/ollama | af4cf55884ac54b9e637cd71dadfe9b7a5685877 | d6786f294519f290b3a5a41d69ea669177491f46 | not found error before pulling model (#718) | [
{
"path": "server/routes.go",
"patch": "@@ -3,8 +3,10 @@ package server\n import (\n \t\"context\"\n \t\"encoding/json\"\n+\t\"errors\"\n \t\"fmt\"\n \t\"io\"\n+\t\"io/fs\"\n \t\"log\"\n \t\"net\"\n \t\"net/http\"\n@@ -170,6 +172,11 @@ func GenerateHandler(c *gin.Context) {\n \n \tmodel, err := GetModel(req... | 2023-10-06T20:06:20 |
huggingface/transformers | b9bd8c45a15b4e90bd800dcf6857cd6d690b7ca4 | baecdb8a97a456ca475c832a11dac1b05ef72e1f | [CI] Build translated docs (#41632)
fix | [
{
"path": ".github/workflows/pr_build_doc_with_comment.yml",
"patch": "@@ -98,7 +98,7 @@ jobs:\n commit_sha: ${{ needs.get-pr-info.outputs.PR_HEAD_SHA }}\n pr_number: ${{ needs.get-pr-number.outputs.PR_NUMBER }}\n package: transformers\n- languages: ar de en es fr hi it ko pt tr zh ja... | 2025-10-16T12:01:33 |
vuejs/vue | 8d885128371c81994715691c81d161b3768706dd | 69771090511fb8072b5600f6c4d212026c8e7eea | ssr: resolve require() calls relative to bundle (fix #4936) | [
{
"path": "package.json",
"patch": "@@ -101,6 +101,7 @@\n \"nightwatch\": \"^0.9.9\",\n \"nightwatch-helpers\": \"^1.2.0\",\n \"phantomjs-prebuilt\": \"^2.1.1\",\n+ \"resolve\": \"^1.2.0\",\n \"rollup\": \"^0.41.4\",\n \"rollup-plugin-alias\": \"^1.2.0\",\n \"rollup-plugin-babel\"... | 2017-02-15T17:07:07 |
denoland/deno | a9b7c7d6d0c452a6697e8b8bbab89f427656c238 | a02d2201797d5e3436a251ff39e4cde6b2d4a13b | fix: enable edns for Deno.resolveDns (#31951)
Currently when using `Deno.resolveDns` on a domain with many (large)
records **and** using a DNS server that allows large replies over UDP,
deno times out with the following error:
```ts
> await Deno.resolveDns('large-txt-records.0x76.dev', 'TXT')
dropped malformed messag... | [
{
"path": "ext/net/01_net.js",
"patch": "@@ -87,7 +87,7 @@ async function resolveDns(query, recordType, options) {\n query,\n recordType,\n options,\n- });\n+ }, /* useEdns0 */ true);\n return ArrayPrototypeMap(res, (recordWithTtl) => recordWithTtl.data);\n } finally {\n if... | 2026-01-26T17:40:56 |
ollama/ollama | 288814d3e4b60a73cdfff80df9df2f572952aca3 | 04733438daa9ce274b3dd1d1cad66b51b43ef6e1 | fix ref counts | [
{
"path": "server/download.go",
"patch": "@@ -36,7 +36,7 @@ type blobDownload struct {\n \n \tdone chan struct{}\n \tcontext.CancelFunc\n-\trefCount atomic.Int32\n+\treferences atomic.Int32\n }\n \n type blobDownloadPart struct {\n@@ -241,8 +241,19 @@ func (b *blobDownload) Write(p []byte) (n int, err error... | 2023-10-03T23:44:35 |
huggingface/transformers | baecdb8a97a456ca475c832a11dac1b05ef72e1f | 44539827d55254546dff5249d976419c798d2f63 | [`Ernie 4.5 Moe`] Fix Moe and offloading (#41385)
fix | [
{
"path": "src/transformers/models/ernie4_5_moe/modeling_ernie4_5_moe.py",
"patch": "@@ -286,37 +286,12 @@ def forward(self, hidden_states):\n return hidden_states + self.e_score_correction_bias.squeeze()\n \n \n-class Ernie4_5_MoeRouter(nn.Module):\n- def __init__(self, config):\n- super(... | 2025-10-16T11:59:01 |
ggml-org/llama.cpp | c7be9febcbafa9af7d1b9443f86475c59c9c5f87 | 8415f61e23d04427cd0d912fbb9d33b85f849456 | [SYCL] fix UT fault cases: count-equal, argsort, pad OPs (#16521)
* fix/refactor OP argsort, pad
* fix count-equal op
* update SYCL OP list
* fix format issue
---------
Co-authored-by: Zhang Jianyu <zhang.jianyu@outlook.com> | [
{
"path": "docs/ops.md",
"patch": "@@ -31,7 +31,7 @@ Legend:\n | CONV_TRANSPOSE_1D | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ |\n | CONV_TRANSPOSE_2D | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |\n | COS | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ❌ |\n-| ... | 2025-10-12T13:53:35 |
vuejs/vue | 69771090511fb8072b5600f6c4d212026c8e7eea | ef57aa250f57f92baa8d7c6c27f96ded20802a00 | improve transition-group move detection (fix #4900, close #4911) | [
{
"path": "src/platforms/web/runtime/components/transition-group.js",
"patch": "@@ -12,13 +12,15 @@\n // nodes will remain where they should be.\n \n import { warn, extend } from 'core/util/index'\n+import { addClass, removeClass } from '../class-util'\n import { transitionProps, extractTransitionData } fro... | 2017-02-15T14:59:43 |
ollama/ollama | 711e891f0f6c2f36b5178c1a65177859eb0855b8 | 090d08422b361bcbef82a04d0e6e160caaad8f89 | fix resumable downloads
glob returns files in lexical order which is not appropriate when
rebuilding the parts list | [
{
"path": "server/download.go",
"patch": "@@ -12,6 +12,7 @@ import (\n \t\"os\"\n \t\"path/filepath\"\n \t\"strconv\"\n+\t\"strings\"\n \t\"sync\"\n \t\"sync/atomic\"\n \t\"time\"\n@@ -39,9 +40,18 @@ type blobDownload struct {\n }\n \n type blobDownloadPart struct {\n+\tN int\n \tOffset int64\n \... | 2023-10-02T22:26:27 |
denoland/deno | a02d2201797d5e3436a251ff39e4cde6b2d4a13b | 11f255c8813bb2463aa33447a5995f82a23ecbcf | feat(node): implement `FileHandle.readv()` method (#31943)
Implements `filehandle.readv(buffers[, position])` method for the
`FileHandle` class in `node:fs/promises`.
This is part of #25554 tracking `FileHandle` missing methods.
References: Node.js docs:
https://nodejs.org/api/fs.html#filehandlereadvbuffers-position
... | [
{
"path": "ext/node/polyfills/_fs/_fs_readv.ts",
"patch": "@@ -18,6 +18,7 @@ import { customPromisifyArgs } from \"ext:deno_node/internal/util.mjs\";\n \n const {\n ObjectDefineProperty,\n+ Promise,\n PromisePrototypeThen,\n TypedArrayPrototypeGetByteLength,\n } = primordials;\n@@ -98,6 +99,11 @@ Obj... | 2026-01-24T14:56:45 |
huggingface/transformers | 143acfe2ceb06b2bc39c2c59412949cfc4c8ed08 | 67fae90519f0992dc27c396d3b112bdf0d004ce5 | fix check inputs for text2text pipeline (#41556)
fix check inputs
Signed-off-by: jiqing-feng <jiqing.feng@intel.com>
Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com> | [
{
"path": "src/transformers/pipelines/text2text_generation.py",
"patch": "@@ -194,17 +194,13 @@ def preprocess(self, inputs, truncation=TruncationStrategy.DO_NOT_TRUNCATE, **kw\n \n def _forward(self, model_inputs, **generate_kwargs):\n in_b, input_length = model_inputs[\"input_ids\"].shape\n-\n... | 2025-10-16T11:42:41 |
ggml-org/llama.cpp | 4b2dae383df708e2afc49c4859a81cd074f5ac10 | 41aac5c69b5fb281bc1f486afb053f78101bb39e | common : update presets (#16504)
* presets : add --embd-gemma-default and remove old embedding presets
* presets : add gpt-oss presets
* presets : add vision presets
* cont : remove reasoning overrides [no ci]
* cont : fix batch size for embedding gemma [no ci] | [
{
"path": "common/arg.cpp",
"patch": "@@ -3358,7 +3358,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex\n add_opt(common_arg(\n {\"--chat-template-kwargs\"}, \"STRING\",\n string_format(\"sets additional params for the json template parser\"),\n- ... | 2025-10-12T06:29:13 |
vuejs/vue | 05c769bf4442d57f4593016d562bb702174ae295 | f59aef08e62576fd574b334510233e0f48198789 | fix .once with other modifiers that prevent execution of a handler (fix #4846) | [
{
"path": "src/compiler/codegen/events.js",
"patch": "@@ -19,11 +19,14 @@ const keyCodes: { [key: string]: number | Array<number> } = {\n const modifierCode: { [key: string]: string } = {\n stop: '$event.stopPropagation();',\n prevent: '$event.preventDefault();',\n- self: 'if($event.target !== $event.c... | 2017-02-14T22:09:37 |
denoland/deno | 11f255c8813bb2463aa33447a5995f82a23ecbcf | a2e3b452504cd1c925b68fae8401103ff0ea0fb2 | feat: add --inspect-publish-uid flag for VSCode debugging (#31927)
This commit adds `--inspect-publish-uid` flag.
This flag is hidden in the help output at the moment. This flag is
getting added because it allows us to properly implement certain debugging
workflows in VSCode. More concretely, the --inspect-publish-ui... | [
{
"path": "cli/args/flags.rs",
"patch": "@@ -53,6 +53,7 @@ use deno_path_util::normalize_path;\n use deno_path_util::resolve_url_or_path;\n use deno_path_util::url_to_file_path;\n use deno_runtime::UnstableFeatureKind;\n+pub use deno_runtime::deno_inspector_server::InspectPublishUid;\n use deno_runtime::den... | 2026-01-24T11:30:46 |
huggingface/transformers | 67fae90519f0992dc27c396d3b112bdf0d004ce5 | af2a66ced94b97e6e435d42be2ac67bc8c980446 | Fix FP-Quant quantization fallback CPU dispatch. (#41619)
* fp_quant fix
* Update quantizer_fp_quant.py | [
{
"path": "src/transformers/quantizers/quantizer_fp_quant.py",
"patch": "@@ -97,6 +97,10 @@ def create_quantized_param(\n ):\n module, _ = get_module_from_name(model, param_name)\n \n+ if target_device == \"cpu\" and param_name.endswith(\"weight\"):\n+ # Works agains hard-coded... | 2025-10-16T11:41:01 |
ggml-org/llama.cpp | 41aac5c69b5fb281bc1f486afb053f78101bb39e | a2fba89a426ff8005d303c73f0436e7e67368b70 | ggml : Fix FP16 ELU positive branch (#16519)
Co-authored-by: Aaron <shelhamer.aaron@gmail.com> | [
{
"path": "ggml/src/ggml-cpu/vec.h",
"patch": "@@ -820,7 +820,8 @@ inline static void ggml_vec_tanh_f16 (const int n, ggml_fp16_t * y, const ggml_f\n inline static void ggml_vec_elu_f32 (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = (x[i] > 0.f) ? x[i] : expm1f(x[i]); }\n in... | 2025-10-12T05:25:37 |
ollama/ollama | d104b7e99772e02da46d55ec9d8d2df724b81112 | 9e2de1bd2c09cfc6a68deb50e7ec5033df6d22ed | Fix go test./... issue: fmt.Println arg list ends with redundant newline (#705) | [
{
"path": "cmd/cmd.go",
"patch": "@@ -687,27 +687,27 @@ func generateInteractive(cmd *cobra.Command, model string) error {\n \t\t\t\tswitch args[1] {\n \t\t\t\tcase \"license\":\n \t\t\t\t\tif resp.License == \"\" {\n-\t\t\t\t\t\tfmt.Println(\"No license was specified for this model.\\n\")\n+\t\t\t\t\t\tfmt... | 2023-10-05T15:11:04 |
vuejs/vue | 0201d8ca1dd9b97d82cf91118aa0478708528989 | 97d36bc23f788974c6d16ba82e98375b1cac96fa | Add foreignObject as a dedicated SVG element (fixes #4813) (#4926)
* Add foreignObject to isSVG list; Fix for #4813
* Remove extra comma, alphabetize | [
{
"path": "src/platforms/web/util/element.js",
"patch": "@@ -25,8 +25,8 @@ export const isHTMLTag = makeMap(\n // this map is intentionally selective, only covering SVG elements that may\n // contain child elements.\n export const isSVG = makeMap(\n- 'svg,animate,circle,clippath,cursor,defs,desc,ellipse,fi... | 2017-02-14T19:36:36 |
denoland/deno | a2e3b452504cd1c925b68fae8401103ff0ea0fb2 | bbc02e4d6ea99049c1c9cee1e0717f042c6fd71c | fix(x): use local npm package even if npm specifier is used, forward unstable flags (#31942)
Fixes #31933.
parts of the changes + tests written by claude, investigation and idea
for the fix + tests by myself | [
{
"path": "cli/args/mod.rs",
"patch": "@@ -1320,6 +1320,16 @@ impl CliOptions {\n unstable_features\n }\n \n+ /// Returns unstable feature flags as CLI arguments (e.g., \"--unstable-unsafe-proto\").\n+ /// This includes features from both CLI flags and config file.\n+ pub fn unstable_args(&self) ->... | 2026-01-24T11:30:15 |
ggml-org/llama.cpp | 20cc625edc2264aae2779e71bef1593e6a4e8c43 | 11f0af5504252e453d57406a935480c909e3ff37 | ggml: Correct SVE implementation in ggml_vec_dot_f16_unroll (#16518)
The previous SVE implementation for `ggml_vec_dot_f16_unroll` contained a bug due to a copy-paste error. The wrong variable was used in an FMA instruction, leading to incorrect results. This commit corrects the variable usage and improves the clarity... | [
{
"path": "ggml/src/ggml-cpu/vec.h",
"patch": "@@ -144,14 +144,14 @@ inline static void ggml_vec_dot_f16_unroll(const int n, const int xs, float * GG\n for (int i = 0; i < np; i += ggml_f16_step) {\n ay1 = GGML_F16x_VEC_LOAD(y + i + 0 * ggml_f16_epr, 0); // 8 elements\n \n- ax... | 2025-10-12T05:15:00 |
vuejs/vue | 01151ce3fa3cb346443d086332ec053c389ee619 | dfaf1265a2367df93ed5dc5484f7578c14174b9f | fix #4872, use context agnostic Function constructor check (#4928)
* fix #4872, use context agnostic Function constructor check
* use getType to check Function Constructor
* fix negation | [
{
"path": "src/core/util/props.js",
"patch": "@@ -70,7 +70,8 @@ function getPropDefaultValue (vm: ?Component, prop: PropOptions, key: string): a\n return vm._props[key]\n }\n // call factory function for non-Function types\n- return typeof def === 'function' && prop.type !== Function\n+ // a value... | 2017-02-14T18:40:57 |
huggingface/transformers | af2a66ced94b97e6e435d42be2ac67bc8c980446 | a59124e27ea9b5a3110495b2c6737c087c8e3bb0 | Migrate transformers cli to Typer (#41487)
* Add typer-slim as explicit dependency
* Migrate CLI to Typer
* code quality
* bump release candidate
* adapt test_cli.py
* Remove ./commands + adapt tests
* fix quality
* consistency
* doctested
* do not serve model in chat
* style
* will it fix them?
* fix test... | [
{
"path": ".gitignore",
"patch": "@@ -98,6 +98,7 @@ celerybeat-schedule\n # Environments\n .env\n .venv\n+.venv*\n env/\n venv/\n ENV/",
"additions": 1,
"deletions": 0,
"language": "Unknown"
},
{
"path": "setup.py",
"patch": "@@ -114,7 +114,7 @@\n \"GitPython<3.1.19\",\n \"hf... | 2025-10-16T11:29:42 |
ggml-org/llama.cpp | a3cb04744fb5c591985f53b749fef5407d07a145 | 4a8fbe0a5eb75f339782ebcf29c17848122184d3 | metal : fix mul-mm condition + fix mul-mv permuted kernels (#16494) | [
{
"path": "ggml/src/ggml-metal/ggml-metal-ops.cpp",
"patch": "@@ -1546,9 +1546,8 @@ int ggml_metal_op_mul_mat(ggml_metal_op_t ctx, int idx) {\n !ggml_is_transposed(op->src[1]) &&\n // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs\n // AMD GPU and older A... | 2025-10-11T13:54:10 |
denoland/deno | bbc02e4d6ea99049c1c9cee1e0717f042c6fd71c | 92e73b939ac0561ababfdd24fc891e517a769f92 | fix(ext/node): `sqlite`'s `StatementSync` compatibility (#31941)
Allows
https://github.com/nodejs/node/blame/v24.12.0/test/parallel/test-sqlite-statement-sync.js
to pass | [
{
"path": "ext/node/ops/sqlite/mod.rs",
"patch": "@@ -98,11 +98,9 @@ pub enum SqliteError {\n #[property(\"code\" = self.code())]\n InvalidExpandedSql,\n #[class(range)]\n- #[error(\n- \"The value of column {0} is too large to be represented as a JavaScript number: {1}\"\n- )]\n+ #[error(\"Value... | 2026-01-24T10:42:14 |
vuejs/vue | 7259aefbe82516c63e9944a4dcc1dce1ac754bff | ebe27295dfee4fee0610563efe3c7fefca0a5d67 | filter division expression with parentheses (fix #4838) (#4844)
* filter division expression with parentheses (fix #4838)
* verify also dot and square brackets
* missed a link check sorry for the useless commit
* added + - and _ support | [
{
"path": "src/compiler/parser/filter-parser.js",
"patch": "@@ -55,7 +55,7 @@ export function parseFilters (exp: string): string {\n p = exp.charAt(j)\n if (p !== ' ') break\n }\n- if (!p || !/[\\w$]/.test(p)) {\n+ if (!p || !/[\\w).\\]\\+\\-\\_$]/.test(p)) {\n ... | 2017-02-14T18:30:11 |
ggml-org/llama.cpp | 31d0ff1869aa2ea31f4e96d5877d0343e9a2171b | 97870e64975b26c5e06a3540a8dc0ff601351e86 | server / ranking : add sorting and management of top_n (#16403)
* server / ranking : add sorting and management of top_n
* Make the retro compatible if no top_n will return
all results
here is a script to make some test
```script
URL=${1:-http://127.0.0.1:8181}
curl "$URL/v1/rerank" -H "Content-Type: application/... | [
{
"path": "tools/server/server.cpp",
"patch": "@@ -5401,15 +5401,6 @@ int main(int argc, char ** argv) {\n \n const json body = json::parse(req.body);\n \n- // TODO: implement\n- //int top_n = 1;\n- //if (body.count(\"top_n\") != 1) {\n- // top_n = body.at(\"top_n\");\... | 2025-10-11T13:39:04 |
huggingface/transformers | c0a5cf19ad2e428722e631cc17d642330faec7ef | 3ef6f2c415ccd4f01460fb3733b6ad709e58d098 | Fix tokenization test (#41649)
fix | [
{
"path": "tests/test_tokenization_common.py",
"patch": "@@ -4356,6 +4356,7 @@ def test_tokenizer_mismatch_warning(self):\n for tokenizer, pretrained_name, kwargs in self.tokenizers_list:\n with self.subTest(f\"{tokenizer.__class__.__name__} ({pretrained_name})\"):\n with... | 2025-10-16T09:14:20 |
ollama/ollama | 0a4f21c0a723e71278e4a619eb263c58d02f256e | 9abb66254a583754fa64757ce8b3270b052ab848 | fix docker build (#659) | [
{
"path": "Dockerfile",
"patch": "@@ -2,7 +2,7 @@ FROM nvidia/cuda:11.8.0-devel-ubuntu22.04\n \n ARG TARGETARCH\n ARG VERSION=0.0.0\n-ARG GOFLAGS=\"'-ldflags -w -s'\"\n+ARG GOFLAGS=\"'-ldflags=-w -s'\"\n \n WORKDIR /go/src/github.com/jmorganca/ollama\n RUN apt-get update && apt-get install -y git build-esse... | 2023-09-30T20:34:01 |
denoland/deno | 92e73b939ac0561ababfdd24fc891e517a769f92 | 59fe2aa577891618d6610a47f8a4ddadd3a39e2a | fix(node): throw ERR_INVALID_ARG_TYPE for invalid fs.readFile path (#31918) | [
{
"path": "ext/node/polyfills/_fs/_fs_readFile.ts",
"patch": "@@ -12,15 +12,18 @@ import {\n import { Buffer } from \"node:buffer\";\n import { readAllSync } from \"ext:deno_io/12_io.js\";\n import { FileHandle } from \"ext:deno_node/internal/fs/handle.ts\";\n-import { pathFromURL } from \"ext:deno_web/00_i... | 2026-01-24T00:36:19 |
ggml-org/llama.cpp | 477a66b03501cf3bd067f8968b77ca4d053ff1bd | e60f01d941bc5b7fae62dd57fee4cec76ec0ea6e | convert : correctly handle LLaMA tokenizer for Jamba (#16470)
* fix: convert_hf_to_gguf - change Jamba non-sentencepiece mode (tokenizer.json) vocab construction
* fix: convert_hf_to_gguf - jamba non-sentencepiece tokenizer to use _set_vocab_llama_hf func
* fix: convert_hf_to_gguf - removed get_vocab_base_pre from j... | [
{
"path": "convert_hf_to_gguf.py",
"patch": "@@ -5966,20 +5966,12 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter\n class JambaModel(TextModel):\n model_arch = gguf.MODEL_ARCH.JAMBA\n \n- def get_vocab_base_pre(self, tokenizer) -> str:\n- del tokenizer # un... | 2025-10-11T08:33:41 |
huggingface/transformers | 3ef6f2c415ccd4f01460fb3733b6ad709e58d098 | 59efd86da2fd2ce9d2af21b68956abf397009ec2 | Allow passing `tp_plan` in `from_pretrained` directly (#41435)
* start
* allow passing it
* fix plans
* fix
* fix
* style
* style
* fix
* add_test
* oupsi indent
* fix
* fix
* fix for CI without accelerator
* fix import | [
{
"path": "src/transformers/integrations/tensor_parallel.py",
"patch": "@@ -38,16 +38,15 @@\n from torch.distributed.tensor import DTensor, Placement, Replicate, Shard\n \n \n-def initialize_tensor_parallelism(tp_plan, tp_size=None, device_mesh=None, device_map=None):\n+def initialize_tensor_parallelism... | 2025-10-16T09:12:07 |
vuejs/vue | f763fc23546c8163676abe24781676ac7f2ca569 | 9df7870b7f25e23384e2053f7bf539f494817096 | small improvements on html-parser (#4932)
* remove unnecessary code
* not need to capture group when decoding attribute value
* fix eslint | [
{
"path": "src/compiler/parser/html-parser.js",
"patch": "@@ -56,8 +56,8 @@ const decodingMap = {\n '&': '&',\n ' ': '\\n'\n }\n-const encodedAttr = /&(lt|gt|quot|amp);/g\n-const encodedAttrWithNewLines = /&(lt|gt|quot|amp|#10);/g\n+const encodedAttr = /&(?:lt|gt|quot|amp);/g\n+const encodedAttr... | 2017-02-14T14:05:00 |
ollama/ollama | 9abb66254a583754fa64757ce8b3270b052ab848 | 1d0ebe67e8754f2cdfc0b986f980a87f71a7efa5 | docker: fix volume permission errors | [
{
"path": "Dockerfile",
"patch": "@@ -15,13 +15,9 @@ RUN /usr/local/go/bin/go generate ./... \\\n && /usr/local/go/bin/go build .\n \n FROM ubuntu:22.04\n-ENV OLLAMA_HOST 0.0.0.0\n-\n RUN apt-get update && apt-get install -y ca-certificates\n-RUN groupadd ollama && useradd -m -g ollama ollama\n-\n COPY ... | 2023-09-30T19:32:06 |
denoland/deno | 59fe2aa577891618d6610a47f8a4ddadd3a39e2a | e6e72e57c2e2566b54f4672e03a6ff4d91d7b167 | fix(ext/web): support object in `DOMException` second argument for Node.js compat (#31939)
Fix #31915 | [
{
"path": "ext/web/01_dom_exception.js",
"patch": "@@ -18,6 +18,7 @@ const {\n ObjectPrototypeIsPrototypeOf,\n ObjectSetPrototypeOf,\n ReflectConstruct,\n+ ReflectHas,\n Symbol,\n SymbolFor,\n } = primordials;\n@@ -95,21 +96,43 @@ class DOMException {\n [_code];\n \n // https://webidl.spec.wh... | 2026-01-23T19:38:56 |
ggml-org/llama.cpp | e60f01d941bc5b7fae62dd57fee4cec76ec0ea6e | 81086cd6a3ca1252f0dc0f938171648399179c53 | server : fix division by zero when reporting stats (#16501) | [
{
"path": "tools/server/server.cpp",
"patch": "@@ -4226,7 +4226,7 @@ struct server_context {\n metrics.on_prompt_eval(slot);\n }\n \n- slot.t_token_generation = (t_current - slot.t_start_generation) / 1e3;\n+ slot.t_token_generation = std::ma... | 2025-10-10T19:15:05 |
huggingface/transformers | 59efd86da2fd2ce9d2af21b68956abf397009ec2 | 7b7d17f9bfbfaa703fbb03aa854a345c20c8aefe | Add aux loss for GLM-4.5V (#41564)
* add aux
* update
* update config to text_config
* use qwen data class to avoid repeat again
* format
* update
* use 1e-4
* update
* update for remove init
* Apply style fixes
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-... | [
{
"path": "src/transformers/models/glm4v_moe/configuration_glm4v_moe.py",
"patch": "@@ -206,7 +206,8 @@ class Glm4vMoeTextConfig(PreTrainedConfig):\n \\--k dense layers--/\n norm_topk_prob (`bool`, *optional*, defaults to `True`):\n... | 2025-10-16T09:04:21 |
ollama/ollama | 9771b1ec510bd0a217a61ef85bb98cabd4818f84 | 76db4a49cfc6e5a600860a7adf05dfe13707542f | windows runner fixes (#637) | [
{
"path": "llm/llama.go",
"patch": "@@ -64,27 +64,29 @@ func chooseRunners(workDir, runnerType string) []ModelRunner {\n \trunnerAvailable := false // if no runner files are found in the embed, this flag will cause a fast fail\n \tfor _, r := range runners {\n \t\t// find all the files in the runner's bin d... | 2023-09-29T15:47:55 |
denoland/deno | 3403492db3e507e22bb16770a48bccfbeb3160fd | 79716029c3d02664be48785d58a413683df9e66f | fix(ext/node): retry stale keepAlive connections in http client (#31932)
Closes https://github.com/denoland/deno/issues/31891
When a server closes an idle keepAlive connection, the client's HTTP
agent may still have the socket pooled.
Attempting to reuse this stale socket results in ECONNRESET ("socket
hang up").
Thi... | [
{
"path": "ext/node/polyfills/http.ts",
"patch": "@@ -432,6 +432,8 @@ class ClientRequest extends OutgoingMessage {\n }\n \n if (this.agent) {\n+ // Store options for potential retry on stale keepAlive connections.\n+ this._agentOptions = optsWithoutSignal;\n this.agent.addRequest(th... | 2026-01-23T15:48:24 |
vuejs/vue | 7e1d07aad240b0d99ce4b62cf26002b39562fbfa | a5a00738b15f24980875709739b45c54d99ee150 | Fix JSFiddle links (#4910)
Use an inline script tag with unpkg to explicitly show what we're
including. It allows to easily test regressions by appending @2.x.x at
the end of the of the link.
It also allows to directly check if the repro is using the latest
version | [
{
"path": ".github/CONTRIBUTING.md",
"patch": "@@ -26,7 +26,7 @@ Hi! I’m really excited that you are interested in contributing to Vue.js. Befo\n \n - Issues with no clear repro steps will not be triaged. If an issue labeled \"need repro\" receives no further input from the issue author for more than 5 days... | 2017-02-12T23:49:47 |
ggml-org/llama.cpp | 68ee98ae181a5c83a5cc6261daeee69a1f588c15 | cdb6da468cc33323955a523738d2e1675aeb5e9a | server : return HTTP 400 if prompt exceeds context length (#16486)
In streaming mode when prompt exceeds context length, the server returns
HTTP 200 status code with a JSON error in the body. This is very
confusing and inconsistent with all other inference engines which return
HTTP 4xx error in this case.
This patch... | [
{
"path": "tools/server/server.cpp",
"patch": "@@ -3727,7 +3727,7 @@ struct server_context {\n }\n } else {\n if (slot.n_prompt_tokens() >= slot.n_ctx) {\n- send_error(slot, \"the request exceeds t... | 2025-10-10T14:11:07 |
ollama/ollama | 92c20fdae643d70a33d1a7e3b34f6a1338cf5e44 | c951da70965981cedcd58827d53eec91d5167e44 | fix error messages for unknown commands in the repl (#611) | [
{
"path": "cmd/cmd.go",
"patch": "@@ -635,6 +635,8 @@ func generateInteractive(cmd *cobra.Command, model string) error {\n \t\t\t\t\t} else {\n \t\t\t\t\t\tusage()\n \t\t\t\t\t}\n+\t\t\t\tdefault:\n+\t\t\t\t\tfmt.Printf(\"Unknown command '/set %s'. Type /? for help\\n\", args[1])\n \t\t\t\t}\n \t\t\t} else ... | 2023-09-28T21:19:45 |
huggingface/transformers | 7b7d17f9bfbfaa703fbb03aa854a345c20c8aefe | e20df45bf676d80bdddb9757eeeafe6c0c81ecfa | 🚨 [v5] Toggle the serialization format in processors (#41474)
* toggle the serialization
* prob this fixes it
* fix tests
* typo
* delete legacy save entirely
* remove extra nesting in if
* revert test and serialzie a public attr instead of private | [
{
"path": "src/transformers/feature_extraction_utils.py",
"patch": "@@ -512,7 +512,10 @@ def get_feature_extractor_dict(\n with open(resolved_feature_extractor_file, encoding=\"utf-8\") as reader:\n text = reader.read()\n feature_extractor_dict = json.loads(text)\n- ... | 2025-10-16T08:19:22 |
vuejs/vue | 1f4cf63e8a7b8c04292207a17bf7f536ca2d08f8 | af619151c6ea9f39ea551770916c1fb6da7f6e41 | tweak template compilation error output | [
{
"path": "src/compiler/error-detector.js",
"patch": "@@ -55,7 +55,7 @@ function checkFor (node: ASTElement, text: string, errors: Array<string>) {\n \n function checkIdentifier (ident: ?string, type: string, text: string, errors: Array<string>) {\n if (typeof ident === 'string' && !identRE.test(ident)) {... | 2017-02-07T00:50:13 |
denoland/deno | e57d3697b2e89f92ed18ac1a985a5697054ff78d | 8eb295d6fb178bceb1ef90d339720cab4d572a32 | fix(ext/node): align assert throws/rejects signatures (#31934)
Fixes #31931 | [
{
"path": "ext/node/polyfills/assert.ts",
"patch": "@@ -519,46 +519,34 @@ function expectsNoError(\n \n function throws(\n fn: () => void,\n- message?: string,\n-): void;\n-function throws(\n- fn: () => void,\n- error?: Function,\n message?: string | Error,\n ): void;\n function throws(\n fn: () =>... | 2026-01-23T07:47:09 |
ollama/ollama | a2b210130f140bc1ec94d80d727e7c3d0e08a1ac | ed20837f9aed15452c29ce3089f1fcec430e814e | fedora install fixes (#609) | [
{
"path": "scripts/install.sh",
"patch": "@@ -216,12 +216,13 @@ fi\n if ! lsmod | grep -q nvidia; then\n KERNEL_RELEASE=\"$(uname -r)\"\n case $OS_NAME in\n- centos|rhel|rocky|fedora|amzn) $SUDO $PACKAGE_MANAGER -y install kernel-devel-$KERNEL_RELEASE kernel-headers-$KERNEL_RELEASE ;;\n+ ... | 2023-09-27T15:43:47 |
huggingface/transformers | 9f71e3a604e329b317e8bb2f659ed2bb66175a56 | bc9900562d20ddb6fefbc0d3d52633ac23787932 | [docs] Duplicate entry (#41591)
fix | [
{
"path": "docs/source/en/_toctree.yml",
"patch": "@@ -544,8 +544,6 @@\n title: Helium\n - local: model_doc/herbert\n title: HerBERT\n- - local: model_doc/hgnet_v2\n- title: HGNet-V2\n - local: model_doc/hunyuan_v1_dense\n title: HunYuanDenseV1\n - local... | 2025-10-15T15:02:36 |
vuejs/vue | a0a619f6119c4cc6c71934c29fe3e62d3d608da9 | bf1f5f18e2c0e5710410ac841583feb6fd99cdec | fix sfc parser hanging on trailing ill-formatted brackets (fix vue-loader/#628) | [
{
"path": "src/compiler/parser/html-parser.js",
"patch": "@@ -168,8 +168,8 @@ export function parseHTML (html, options) {\n parseEndTag(stackedTag, index - endTagLength, index)\n }\n \n- if (html === last && options.chars) {\n- options.chars(html)\n+ if (html === last) {\n+ options... | 2017-02-06T23:24:13 |
denoland/deno | 2c0ee56e42b81dbb62aca8211f2d8c0e796f1915 | a8d4565575e23f2297e5c044aa362d3c2143840b | fix(ext/node): implement `sqlite`'s missing options (#31919)
Closes #31634 | [
{
"path": "ext/node/ops/sqlite/database.rs",
"patch": "@@ -48,6 +48,10 @@ struct DatabaseSyncOptions {\n read_only: bool,\n allow_extension: bool,\n enable_double_quoted_string_literals: bool,\n+ use_big_int_arguments: bool,\n+ allow_bare_named_params: bool,\n+ return_arrays: bool,\n+ allow_unknow... | 2026-01-23T04:01:45 |
huggingface/transformers | bc9900562d20ddb6fefbc0d3d52633ac23787932 | 72fd67929b9a877cb2934985629d80b158f37689 | Fix quantization base class (#41613)
* fix
* fix
---------
Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com> | [
{
"path": "src/transformers/quantizers/base.py",
"patch": "@@ -211,6 +211,9 @@ def update_ep_plan(self, config):\n \"updates the tp plan for the scales\"\n return config\n \n+ def _process_model_before_weight_loading(self, model, **kwargs):\n+ return model\n+\n def preprocess_m... | 2025-10-15T14:58:17 |
vuejs/vue | f968e6ba3a1b22cce51a924b055f030d29017f6a | ec7fca8495528bda169c44ac660747bf2a8000a8 | fix v-pre end check for unary tags | [
{
"path": "src/compiler/parser/index.js",
"patch": "@@ -52,13 +52,25 @@ export function parse (\n transforms = pluckModuleFunction(options.modules, 'transformNode')\n postTransforms = pluckModuleFunction(options.modules, 'postTransformNode')\n delimiters = options.delimiters\n+\n const stack = []\n ... | 2017-02-06T23:00:43 |
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.