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
denoland/deno
e55b986c1c895e44f23be5b388b52460b2591088
4b0fdb1ca0ff8fbefd30b9e3a3f93e3593cc4b8e
fix(ext/node): handle generateKeyPair 2-arg call and enable PQC tests (#32619) ## Summary - `generateKeyPair(type, callback)` without options was throwing `ERR_INVALID_ARG_TYPE` because the callback was being treated as the options argument. Now properly detects when the second argument is a function and shifts ...
[ { "path": "ext/node/polyfills/internal/crypto/keygen.ts", "patch": "@@ -568,16 +568,20 @@ export function generateKeyPair(\n export function generateKeyPair(\n type: KeyType,\n options: unknown,\n- callback: (\n+ callback?: (\n err: Error | null,\n publicKey: any,\n privateKey: any,\n ) ...
2026-03-11T08:01:10
vuejs/vue
05299610ea3e89ddbcfe4d8ede0c298223766423
6ee684983b1f3384a4050d7c47cee7c6a325db8b
fix(vdom): svg inside foreignObject should be rendered with correct namespace (fix #7330) (#7350) * add failed test case * fix failed test case * fix(vdom): svg inside foreignObject should be rendered with correct namespace * adjust comments
[ { "path": "src/core/vdom/create-element.js", "patch": "@@ -139,7 +139,8 @@ function applyNS (vnode, ns, force) {\n if (isDef(vnode.children)) {\n for (let i = 0, l = vnode.children.length; i < l; i++) {\n const child = vnode.children[i]\n- if (isDef(child.tag) && (isUndef(child.ns) || isTru...
2018-01-05T22:44:44
huggingface/transformers
cba7ae865b16f85b6008aebb220461a8863075f4
629c0da46d10d6cc9101d5042c0cdf670d1299e2
fix the FSDP1 default for reshard_after_forward (#42578) Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
[ { "path": "src/transformers/training_args.py", "patch": "@@ -2755,7 +2755,7 @@ def _process_fsdp_args(self):\n )\n fsdp_plugin_args[\"backward_prefetch\"] = prefetch_policy.upper()\n fsdp_plugin_args[\"reshard_after_forward\"] = str(\n- self...
2025-12-03T13:18:20
ggml-org/llama.cpp
acec774ef619c81d4c59adca0e548b3ba035554f
5c0d18881e0e9794c96b2602736b758bac9d9388
HIP: Refactor mma for RDNA and CDNA (#17990) * mma.cuh for rdna4 * mma for rdna3 * mmq for rdna4 * mmq for rdna3 * align i-major and j-major * cdna * fix cuda error * add missing tile of mfma * fix j-major wrong ne on CDNA * fix gramma and empty spaces --------- Co-authored-by: zhang hui <you@example.com>
[ { "path": "ggml/src/ggml-cuda/mma.cuh", "patch": "@@ -76,15 +76,31 @@ namespace ggml_cuda_mma {\n // For the A/C matrices this means I major == row major, J major == column major.\n // For the B matrix this means I major == column major, J major == row major.\n // MIRRORED == Each da...
2025-12-17T08:34:54
denoland/deno
bd3fd9a8d0113bba467d506b00ce839c3bcb8e83
130a3771c5dd0820e0c399c5df9cfd8922b7b6cd
fix(ext/http): allow rejecting HTTP upgrades with non-101 status codes (#32615) Previously, when using node:http's upgrade event, attempting to reject an upgrade request with a non-101 status code (e.g., 401 Unauthorized) would throw a "write UNKNOWN" error because the UpgradeStream parser only accepted 101 Switching ...
[ { "path": "ext/http/http_next.rs", "patch": "@@ -1368,13 +1368,19 @@ enum UpgradeStreamWriteState {\n AsyncMut<Option<(NetworkStreamReadHalf, Bytes)>>,\n ),\n Network(NetworkStreamWriteHalf),\n+ /// The upgrade was rejected with a non-101 status code.\n+ /// The response has been sent and the stre...
2026-03-11T07:28:07
vuejs/vue
6ee684983b1f3384a4050d7c47cee7c6a325db8b
5b4e6679cbfabd7a6229e86f10c300bcc1b2f815
fix(types): contravariant generic default in ComponentOption (#7369)
[ { "path": "types/options.d.ts", "patch": "@@ -6,16 +6,17 @@ type Constructor = {\n }\n \n // we don't support infer props in async component\n-export type Component<Data=DefaultData<Vue>, Methods=DefaultMethods<Vue>, Computed=DefaultComputed, Props=DefaultProps> =\n+// N.B. ComponentOptions<V> is contravari...
2018-01-03T06:33:48
ollama/ollama
40bc4622ef837627034e7ad0d64b646c7bfbf07a
c0f818a07a6ee18105855aafa29e9d8178ed1733
Fix exe name for zip packaging on windows The zip file encodes the OS and architecture, so keep the short exe name
[ { "path": "scripts/build_windows.ps1", "patch": "@@ -82,7 +82,7 @@ function buildOllama() {\n if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}\n }\n New-Item -ItemType Directory -Path .\\dist\\windows-amd64\\ -Force\n- cp .\\ollama.exe .\\dist\\windows-amd64\\ollama-windows-amd64.exe\n+ ...
2024-04-26T16:16:53
huggingface/transformers
629c0da46d10d6cc9101d5042c0cdf670d1299e2
75c135d76f6db226d83213c6bb658a474dddb07d
Fix loaded data order bug when resuming from epoch >= 1 (#40691) * fix resume from epoch >= 1 * add test checking order of sampled data points * add require_torch_non_multi_accelerator decorator to test method * move the epoch setting of epoch_dataloader before iterating over it * make fixup
[ { "path": "src/transformers/trainer.py", "patch": "@@ -2428,8 +2428,6 @@ def _inner_training_loop(\n \n for epoch in range(epochs_trained, num_train_epochs):\n epoch_dataloader = train_dataloader\n- if hasattr(epoch_dataloader, \"set_epoch\"):\n- epoch_dataloade...
2025-12-03T13:14:02
vuejs/vue
400aa651ba1648185274c10aa83c94cb9d8eac56
6be9fdbc8374225b98620cae6d87d97d861913cc
chore: fix setup script (#7364)
[ { "path": "scripts/setup.js", "patch": "@@ -2,9 +2,9 @@ const { test, ln, chmod } = require('shelljs')\n \n function installHooks () {\n if (test('-e', '.git/hooks')) {\n- ln('-sf', '../../build/git-hooks/pre-commit', '.git/hooks/pre-commit')\n+ ln('-sf', '../../scripts/git-hooks/pre-commit', '.git/...
2018-01-02T16:21:05
denoland/deno
130a3771c5dd0820e0c399c5df9cfd8922b7b6cd
b7cc30f2c4c5f581022cb5c55829c741a2862dd5
fix(ext/node): fix multiple node:dgram compatibility issues (#32520) ## Summary Fixes several `node:dgram` compatibility issues, enabling 7 additional node_compat tests: - **`close()` on already-closed socket**: No longer throws `ERR_SOCKET_DGRAM_NOT_RUNNING` — returns early silently (fixes `test-dgram-abort-closed`...
[ { "path": "ext/node/Cargo.toml", "patch": "@@ -72,6 +72,6 @@ errno = \"0.3.10\"\n nix = { workspace = true, features = [\"user\"] }\n \n [target.'cfg(windows)'.dependencies]\n-windows-sys = { workspace = true, features = [\"Win32_Networking_WinSock\"] }\n+windows-sys = { workspace = true, features = [\"Win3...
2026-03-11T07:04:49
ggml-org/llama.cpp
5c0d18881e0e9794c96b2602736b758bac9d9388
4b2a4778f81f222c12271ce3b1997990b3071faf
llama.android : Rewrite Android binding (w/o cpu_features dep) (#17413) * UI: implement basic UI components * util: implement performance monitor; wrap it with a viewmodel * util: implement user preferences utility * UI: implement core flow's screens * UI: add a new MainActivity; update manifest * [WIP] DI: imple...
[ { "path": "CODEOWNERS", "patch": "@@ -32,7 +32,7 @@\n /examples/export-docs/ @ggerganov\n /examples/gen-docs/ @ggerganov\n /examples/gguf/ @ggerganov\n-/examples/llama.android/ @ggerganov\n+/examples/llama.android/ @g...
2025-12-17T08:14:47
huggingface/transformers
75c135d76f6db226d83213c6bb658a474dddb07d
232ecf2cd82c7d3a0bcb13e181d3a10f87eb9fa7
[XPU] Fix fp8 UT patch func (#42584) * [XPU] Fix fp8 UT patch * make style --------- Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>
[ { "path": "tests/quantization/finegrained_fp8/test_fp8.py", "patch": "@@ -46,6 +46,9 @@ def _patch_no_accelerator():\n stack.enter_context(patch(\"torch.cuda.is_available\", return_value=False))\n if hasattr(torch, \"xpu\"):\n stack.enter_context(patch(\"torch.xpu.is_available\",...
2025-12-03T12:22:54
vuejs/vue
f8cb3a239332804808ca058d45609a860dbda55a
dccd182b6763d8ef1871949029c85495ca958246
chore: fix lint
[ { "path": "test/unit/modules/compiler/compiler-options.spec.js", "patch": "@@ -9,58 +9,58 @@ describe('compile options', () => {\n <input type=\"text\" v-model=\"msg\" required max=\"8\" v-validate:field1.group1.group2>\n </div>\n `, {\n- directives: {\n- validate (el, dir) {\n...
2017-12-25T16:05:54
denoland/deno
ed367f222d10d1cad82b5aef9d85498a1289a995
042a768744ef8bfe237540d61336471004755948
fix(npm): memoize peer cache hit checks to prevent combinatorial explosion (#32609) ## Summary - Fixes `deno add npm:@aws-cdk/aws-ecs` (and similar large peer dep trees) hanging indefinitely - Root cause: `find_peers_cache_hit_inner` in `libs/npm/resolution/graph.rs` recursively checks whether cached peer resolutions...
[ { "path": "libs/npm/resolution/graph.rs", "patch": "@@ -2298,18 +2298,33 @@ impl<'a, TNpmRegistryApi: NpmRegistryApi>\n parent_pkgs: &BTreeMap<StackString, NodeId>,\n ) -> Option<PeersResolution> {\n let mut checking = HashSet::new();\n- self.find_peers_cache_hit_inner(nv, parent_pkgs, &mut che...
2026-03-10T17:50:55
ggml-org/llama.cpp
0e49a7b8b47aaa8995953ee7e3c0b5c39ead19fa
4164596c76cb4f62322fd1ff9d157cee5ba97b23
llama-fit-params: fix underflow for dense models (#18095)
[ { "path": "src/llama.cpp", "patch": "@@ -481,8 +481,13 @@ static void llama_params_fit_impl(\n } else {\n LLAMA_LOG_INFO(\"%s: filling dense-only layers back-to-front:\\n\", __func__);\n }\n- uint32_t n_unassigned = hp_ngl;\n for (int id = nd - 1; id >= 0; id--) {\n+ uint32_t n...
2025-12-16T23:47:37
huggingface/transformers
15b79ea808237691397be558bdebbef378b9fa35
377a8ee73f210476c4efb15170d0c32ad3b2c653
[Quantization] fix fbgemm (#42561) * initial commit * passing tests * fix replace_linear * style * rm list * fix * style
[ { "path": "src/transformers/integrations/fbgemm_fp8.py", "patch": "@@ -12,7 +12,11 @@\n # See the License for the specific language governing permissions and\n # limitations under the License.\n \n+from typing import Optional\n+\n from ..activations import ACT2FN\n+from ..core_model_loading import Conversio...
2025-12-03T09:14:28
vuejs/vue
dccd182b6763d8ef1871949029c85495ca958246
44420b7bbcd9f6f4b0df3263e19ad19a974fa7f1
fix: allow codebase to be inlined directly in HTML (#7314) escape - to avoid regex being parsed as HTML comment when entire codebase is inlined. fix #7298
[ { "path": "src/compiler/parser/html-parser.js", "patch": "@@ -22,7 +22,8 @@ const startTagOpen = new RegExp(`^<${qnameCapture}`)\n const startTagClose = /^\\s*(\\/?)>/\n const endTag = new RegExp(`^<\\\\/${qnameCapture}[^>]*>`)\n const doctype = /^<!DOCTYPE [^>]+>/i\n-const comment = /^<!--/\n+// #7298: esc...
2017-12-25T15:50:36
denoland/deno
042a768744ef8bfe237540d61336471004755948
527c844f9574895feff910c5cd2195407422dcc1
fix(ext/node): add `node_api_create_property_key_(latin1/utf8)` (#32559) Implements `node_api_create_property_key_latin1` and `node_api_create_property_key_utf8` to complement the existing `node_api_create_property_key_utf16`. These create internalized V8 strings for use as object property keys, enabling deduplication...
[ { "path": "ext/napi/generated_symbol_exports_list_linux.def", "patch": "@@ -1 +1 @@\n-{ \"node_api_create_syntax_error\"; \"napi_make_callback\"; \"napi_has_named_property\"; \"napi_async_destroy\"; \"napi_coerce_to_object\"; \"napi_get_arraybuffer_info\"; \"napi_detach_arraybuffer\"; \"napi_get_undefined\"...
2026-03-10T16:23:00
ollama/ollama
8589d752ac9c9c5fd81be6179f3bf77243c145f2
de4ded68b0d624d8e37741e08f7b136d98937406
Fix release CI download-artifact path was being used incorrectly. It is where to extract the zip not the files in the zip to extract. Default is workspace dir which is what we want, so omit it
[ { "path": ".github/workflows/release.yaml", "patch": "@@ -311,29 +311,18 @@ jobs:\n - uses: actions/download-artifact@v4\n with:\n name: generate-windows-cpu\n- path: |\n- llm/build\n- dist/windows-amd64\n - uses: actions/download-artifact@v4\n ...
2024-04-26T00:27:11
huggingface/transformers
51c5a7a658b0d56f418d491c4a77073e113dce12
a5c061d24e31706c18670c0ae6579b5c72d545f3
[Quantization] per tensor quantization kernel (#42560) * fix * style * initial * fix * comment * style * fix
[ { "path": "src/transformers/integrations/finegrained_fp8.py", "patch": "@@ -156,6 +156,79 @@ def _w8a8_block_fp8_matmul(\n tl.store(c_ptrs, c, mask=c_mask)\n \n \n+@triton.jit\n+def _w8a8_block_fp8_matmul_per_tensor(\n+ # Pointers to inputs and output\n+ A,\n+ B,\n+ C,\n+ As,\n+ Bs,\n+...
2025-12-02T17:57:16
vuejs/vue
60fb8143f1235d0ae19d56939c424ad46bb33e61
c05bbdfc36d0a2777c5721932a06f098860536b9
chore: fix OpenCollective links
[ { "path": "BACKERS.md", "patch": "@@ -43,8 +43,8 @@ Funds donated via Patreon goes directly to support Evan You's full-time work on\n \n <h2 align=\"center\">Platinum via OpenCollective</h2>\n \n-<a href=\"https://opencollective.com/vuejs/tiers/platinumsponsors/0/website\" target=\"_blank\"><img src=\"https...
2017-12-22T19:48:21
ggml-org/llama.cpp
4164596c76cb4f62322fd1ff9d157cee5ba97b23
ef83fb8601229ff650d952985be47e82d644bfaa
llama-fit-params: QoL impr. for prints/errors (#18089)
[ { "path": "tools/fit-params/fit-params.cpp", "patch": "@@ -4,7 +4,11 @@\n #include \"common.h\"\n #include \"log.h\"\n \n-#include <iostream>\n+#include <chrono>\n+#include <cinttypes>\n+#include <thread>\n+\n+using namespace std::chrono_literals;\n \n #if defined(_MSC_VER)\n #pragma warning(disable: 4244 4...
2025-12-16T23:03:19
denoland/deno
527c844f9574895feff910c5cd2195407422dcc1
ed9858527a2b544a802f7f441b5074f59a1c1284
fix: flush CPU profile and coverage data on Deno.exit() (#32591) ## Summary - When `Deno.exit()` is called, `std::process::exit()` bypasses normal cleanup in the worker's run loop, causing CPU profile and coverage data to be lost - Adds `OpExitCallbacks` stored in `OpState` that are invoked by `op_exit` before `std::...
[ { "path": "cli/worker.rs", "patch": "@@ -23,6 +23,7 @@ use deno_runtime::WorkerExecutionMode;\n use deno_runtime::coverage::CoverageCollector;\n use deno_runtime::cpu_prof_filename;\n use deno_runtime::cpu_profiler::CpuProfiler;\n+use deno_runtime::deno_os::OpExitCallbacks;\n use deno_runtime::deno_permissi...
2026-03-10T16:04:41
ollama/ollama
00b0699c75fc99b998f3394c04e5a16aa4c49eab
993cf8bf55745a46ce756461008bbea3ad8e9cb1
Reload model if `num_gpu` changes (#3920) * reload model if `num_gpu` changes * dont reload on -1 * fix tests
[ { "path": "server/sched.go", "patch": "@@ -421,23 +421,29 @@ func (runner *runnerRef) needsReload(ctx context.Context, req *LlmRequest) bool\n \tslog.Debug(\"evaluating already loaded\", \"model\", req.model.ModelPath)\n \trunner.refMu.Lock()\n \tdefer runner.refMu.Unlock()\n-\t// Ignore the NumGPU settings...
2024-04-25T23:02:40
huggingface/transformers
a5c061d24e31706c18670c0ae6579b5c72d545f3
675e8763cafd74894ce4699d7583bd387beac14e
[`CI`] Fix copies (#42571) * fix * fix circular condition
[ { "path": "src/transformers/models/fast_vlm/modeling_fast_vlm.py", "patch": "@@ -59,7 +59,7 @@ def forward(self, image_features):\n class FastVlmPreTrainedModel(PreTrainedModel):\n config: FastVlmConfig\n base_model_prefix = \"model\"\n- input_modalities = [\"image\", \"text\"]\n+ input_modali...
2025-12-02T17:41:09
vuejs/vue
c05bbdfc36d0a2777c5721932a06f098860536b9
144a4dd860b20ca48263bac150286f627e08d953
chore: fixes gold / platinum sponsors on open collective (#7311) Fixes https://github.com/opencollective/opencollective/issues/777
[ { "path": "README.md", "patch": "@@ -143,16 +143,16 @@ Funds donated via Patreon goes directly to support Evan You's full-time work on\n \n <h4 align=\"center\">Platinum</h4>\n \n-<a href=\"https://opencollective.com/vuejs/tiers/platinumsponsors/0/website\" target=\"_blank\"><img src=\"https://opencollectiv...
2017-12-22T19:46:01
ggml-org/llama.cpp
ef83fb8601229ff650d952985be47e82d644bfaa
ec98e20021f7611db3bbcf6bb6629fed6e1ce4f0
model: fix LFM2 missing tensors (#18105)
[ { "path": "src/llama-model.cpp", "patch": "@@ -6236,8 +6236,8 @@ bool llama_model::load_tensors(llama_model_loader & ml) {\n {\n tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, \"weight\"), {n_embd, n_vocab}, 0);\n \n- output_norm = create_tensor(tn(...
2025-12-16T18:07:43
denoland/deno
ed9858527a2b544a802f7f441b5074f59a1c1284
9c73df383e00295818fa6e77da527a2cd09fb696
fix(npm): support npmrc email authorization (#32616)
[ { "path": "libs/npm_cache/remote.rs", "patch": "@@ -7,7 +7,9 @@ use deno_npm::npm_rc::RegistryConfig;\n #[derive(Debug, thiserror::Error, deno_error::JsError)]\n pub enum AuthHeaderForNpmRegistryError {\n #[class(type)]\n- #[error(\"Both the username and password must be provided for basic auth\")]\n+ #...
2026-03-10T16:03:16
huggingface/transformers
675e8763cafd74894ce4699d7583bd387beac14e
80b408d15fdc0766dbb83a5d0b446891e5fe6549
Allow fallback to loading from Auto"SubProcessor".from_pretrained when model_type can't be inferred from config (#42402) * fix raise error early * add back feature extractor saving logic
[ { "path": "src/transformers/models/auto/processing_auto.py", "patch": "@@ -317,7 +317,6 @@ def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):\n processor_class = config_dict.get(\"processor_class\", None)\n if \"AutoProcessor\" in config_dict.get(\"aut...
2025-12-02T17:26:41
vuejs/vue
d8b08387a293c99b95c1efcf2517447335a618db
956756b1be7084daf8b6afb92ac0da7c24cde2a5
fix: clean up custom events when patched component no longer have events fix #7294
[ { "path": "src/core/instance/lifecycle.js", "patch": "@@ -257,11 +257,10 @@ export function updateChildComponent (\n }\n \n // update listeners\n- if (listeners) {\n- const oldListeners = vm.$options._parentListeners\n- vm.$options._parentListeners = listeners\n- updateComponentListeners(vm, l...
2017-12-21T15:36:03
ggml-org/llama.cpp
ec98e20021f7611db3bbcf6bb6629fed6e1ce4f0
59977eba7b0a3603d0017717d3beec7bde018f3c
llama: fix early stop in params_fit if ctx is set (#18070)
[ { "path": "src/llama.cpp", "patch": "@@ -241,6 +241,13 @@ static void llama_params_fit_impl(\n global_surplus += memory_reduction;\n LLAMA_LOG_INFO(\"%s: context size reduced from %\" PRIu32 \" to %\" PRIu32 \" -> need %\" PRId64 \" MiB less memory in total\\n\",\n ...
2025-12-16T13:24:00
denoland/deno
88d6de0e5763a03eb7237df731590c5864638ecb
f81e36050aa75996ca989638e3dc07de05154de2
fix(ext/node): run worker_threads eval code in sloppy mode (#32428) ## Summary - **Root cause**: `node:worker_threads` with `{ eval: true }` was wrapping code in `import` statements and loading it as a `data:text/javascript` URL via `load_main_es_module`, which enforces **strict mode**. Node.js runs eval workers as C...
[ { "path": "ext/node/polyfills/worker_threads.ts", "patch": "@@ -51,7 +51,6 @@ import { createRequire } from \"node:module\";\n \n const {\n ArrayIsArray,\n- encodeURIComponent,\n Error,\n FunctionPrototypeCall,\n NumberIsFinite,\n@@ -359,24 +358,29 @@ class NodeWorker extends EventEmitter {\n ...
2026-03-10T14:44:53
huggingface/transformers
80b408d15fdc0766dbb83a5d0b446891e5fe6549
52b988d83d8737f25b714128bf6f015338f629a1
[SAM3] Compute masks once instead of per-layer, fix fa2 crash (#42543) * Compute masks once instead of per-layer, fix fa2 crash * nit * Change after review
[ { "path": "src/transformers/models/sam3/modeling_sam3.py", "patch": "@@ -14,9 +14,8 @@\n # limitations under the License.\n \n \n-import collections.abc\n import math\n-from collections.abc import Callable\n+from collections.abc import Callable, Iterable\n from dataclasses import dataclass\n from typing imp...
2025-12-02T17:23:29
ggml-org/llama.cpp
59977eba7b0a3603d0017717d3beec7bde018f3c
79dbae034afdfaa8e17989ea1b9b20094c8d0a36
server: fix crash when batch > ubatch with embeddings (#17912) * server: fix crash when batch > ubatch with embeddings (#12836) Fixes #12836 where the server crashes with GGML_ASSERT failure when running with embeddings enabled and n_batch > n_ubatch. Root cause: Embeddings use non-causal attention which requires al...
[ { "path": "tools/server/server.cpp", "patch": "@@ -73,8 +73,18 @@ int main(int argc, char ** argv, char ** envp) {\n return 1;\n }\n \n+ // validate batch size for embeddings\n+ // embeddings require all tokens to be processed in a single ubatch\n+ // see https://github.com/ggml-org/lla...
2025-12-16T12:27:36
vuejs/vue
956756b1be7084daf8b6afb92ac0da7c24cde2a5
8335217cb4bd13fb07e08a76c07df0fceed6c197
refactor: use more efficient on-demand clone to handle reused node edge cases removes unnecessary slot/static node clones, fix #7292
[ { "path": "src/core/instance/lifecycle.js", "patch": "@@ -239,7 +239,7 @@ export function updateChildComponent (\n // update $attrs and $listeners hash\n // these are also reactive so they may trigger child update if the child\n // used them during render\n- vm.$attrs = (parentVnode.data && parentVno...
2017-12-21T00:44:24
ollama/ollama
14476d48ccfcff46e3b1b5beb9372e8265da5255
ce8ce8256728d4211e50e0db72195073ad37127c
fixes for gguf (#3863)
[ { "path": "llm/gguf.go", "patch": "@@ -190,8 +190,6 @@ func (llm *gguf) Decode(rs io.ReadSeeker) error {\n \t\tllm.kv[k] = v\n \t}\n \n-\tslog.Debug(fmt.Sprintf(\"general.architecture = %s\", llm.kv[\"general.architecture\"]))\n-\n \t// decode tensors\n \tfor i := 0; uint64(i) < llm.numTensor(); i++ {\n \t\...
2024-04-24T03:57:20
ggml-org/llama.cpp
7f2b2f3c778f430edc57d5728641317e9ac5a505
7b1db3d3b770d0affbf3aadee033e1614280085f
arch: refactor LLM_TENSOR_NAMES (#18051) * arch: refactor LLM_TENSOR_NAMES * update docs * typo * fix LLM_ARCH_NEMOTRON_H_MOE * show more meaningful error message on missing tensor * fix and tested LLM_ARCH_NEMOTRON_H_MOE
[ { "path": "docs/development/HOWTO-add-model.md", "patch": "@@ -97,7 +97,7 @@ The model params and tensors layout must be defined in `llama.cpp` source files:\n 1. Define a new `llm_arch` enum value in `src/llama-arch.h`.\n 2. In `src/llama-arch.cpp`:\n - Add the architecture name to the `LLM_ARCH_NAMES`...
2025-12-16T12:22:30
denoland/deno
f81e36050aa75996ca989638e3dc07de05154de2
4c548d42e5895eb3a6f00bb6d7cb3a9d601db7e8
fix(lsp): tsgo scope attribution for asset files (#32608)
[ { "path": "cli/lsp/documents.rs", "patch": "@@ -1159,7 +1159,7 @@ pub struct DocumentModules {\n dep_info_by_scope: once_cell::sync::OnceCell<Arc<DepInfoByScope>>,\n modules_unscoped: Arc<WeakDocumentModuleMap>,\n modules_by_scope: Arc<BTreeMap<Arc<Url>, Arc<WeakDocumentModuleMap>>>,\n- assigned_scop...
2026-03-10T14:43:29
vuejs/vue
8335217cb4bd13fb07e08a76c07df0fceed6c197
49aae6bb157e0650507974b7a9a1b0f2215e400b
fix(core): handle edge cases for functional component returning arrays fix #7282
[ { "path": "src/core/vdom/create-component.js", "patch": "@@ -107,7 +107,7 @@ export function createComponent (\n context: Component,\n children: ?Array<VNode>,\n tag?: string\n-): VNode | void {\n+): VNode | Array<VNode> | void {\n if (isUndef(Ctor)) {\n return\n }", "additions": 1, "d...
2017-12-20T16:02:42
ggml-org/llama.cpp
fb644247de14c616b10deb5e6b17e6f4230f0601
5f5f9b46376ac14d7f95b0d968c182f522602880
CLI: fixed adding cli and completion into docker containers, improved docs (#18003) Co-authored-by: Andrew Aladjev <andrew.aladjev@gmail.com>
[ { "path": ".devops/cann.Dockerfile", "patch": "@@ -107,7 +107,7 @@ ENTRYPOINT [\"/app/tools.sh\"]\n # ENTRYPOINT [\"/app/llama-server\"]\n \n ### Target: light\n-# Lightweight image containing only llama-cli\n+# Lightweight image containing only llama-cli and llama-completion\n # ===========================...
2025-12-16T10:52:23
ollama/ollama
5445aaa94e6f4c5626222177ed53ebd07225dd9a
2ac3dd6853a45237ac049d0a4982becf91ca8c45
Add back memory escape valve If we get our predictions wrong, this can be used to set a lower memory limit as a workaround. Recent multi-gpu refactoring accidentally removed it, so this adds it back.
[ { "path": "llm/memory.go", "patch": "@@ -3,6 +3,8 @@ package llm\n import (\n \t\"fmt\"\n \t\"log/slog\"\n+\t\"os\"\n+\t\"strconv\"\n \t\"strings\"\n \n \t\"github.com/ollama/ollama/api\"\n@@ -49,6 +51,17 @@ func EstimateGPULayers(gpus []gpu.GpuInfo, ggml *GGML, projectors []string, opts\n \tfor _, info := ...
2024-04-24T00:09:02
denoland/deno
4c548d42e5895eb3a6f00bb6d7cb3a9d601db7e8
f33915b81c39ac23a4612ade19310b4611b8b0c1
fix(ext/napi): run async work execute callback on a worker thread (#32560) ## Summary - **Fixes `napi_queue_async_work`** to run the `execute` callback on a worker thread (via `std::thread::spawn`) instead of the V8 main thread - **Dispatches `complete`** back to the main thread via `async_work_sender.spawn()` wh...
[ { "path": "ext/napi/node_api.rs", "patch": "@@ -586,8 +586,16 @@ pub(crate) fn napi_queue_async_work(\n }\n \n let work = SendPtr(work);\n+ let sender = env.async_work_sender.clone();\n+ let tracker = env.external_ops_tracker.clone();\n \n- env.add_async_work(move || {\n+ // Keep the event loop aliv...
2026-03-10T14:19:29
vuejs/vue
49aae6bb157e0650507974b7a9a1b0f2215e400b
e055df82fec0e76e4bc65e5a265b42e208595430
fix(types): make VNodeChildrenArrayContents type more accurate (#7287)
[ { "path": "types/test/options-test.ts", "patch": "@@ -1,4 +1,4 @@\n-import Vue from \"../index\";\n+import Vue, { VNode } from \"../index\";\n import { AsyncComponent, ComponentOptions, FunctionalComponentOptions } from \"../index\";\n import { CreateElement } from \"../vue\";\n \n@@ -277,6 +277,19 @@ Vue.c...
2017-12-20T14:09:23
ggml-org/llama.cpp
5f5f9b46376ac14d7f95b0d968c182f522602880
3d86c6c2b546ae5ce9cdee3ca0ac4a8181e0a073
server: Update README.md incorrect argument (#18073) n-gpu-layer is incorrect argument is n-gpu-layers with the 's'
[ { "path": "tools/server/README.md", "patch": "@@ -1430,7 +1430,7 @@ Model presets allow advanced users to define custom configurations using an `.in\n llama-server --models-preset ./my-models.ini\n ```\n \n-Each section in the file defines a new preset. Keys within a section correspond to command-line argum...
2025-12-16T10:50:43
vuejs/vue
e055df82fec0e76e4bc65e5a265b42e208595430
7cc0b559e9e57fcb3baeae5d8d4c8964aa335b5e
fix(weex): default value for editor, fix #7165 (#7286) * fix(weex): default value for editor, fix #7165 /cc Hanks10100 * fix(weex): recycle-list test
[ { "path": "test/weex/cases/recycle-list/components/editor.vue", "patch": "@@ -10,7 +10,7 @@\n props: ['message'],\n data () {\n return {\n- output: this.message | ''\n+ output: this.message || ''\n }\n }\n }", "additions": 1, "deletions": 1, "language": "U...
2017-12-20T14:05:21
huggingface/transformers
52b988d83d8737f25b714128bf6f015338f629a1
3f17410925a0c62f6844f1395270bf0066133cbf
Transformers serve fix (#42570) * Fix: lacking EOS token + failing AutoProcessor * Tests * Tests
[ { "path": "src/transformers/cli/serve.py", "patch": "@@ -36,7 +36,7 @@\n from tqdm import tqdm\n \n import transformers\n-from transformers import BitsAndBytesConfig, GenerationConfig\n+from transformers import AutoTokenizer, BitsAndBytesConfig, GenerationConfig, PreTrainedTokenizerBase\n from transformers....
2025-12-02T17:19:36
denoland/deno
1036b2ac9046d69a3a8ef7ac30298363a4acc432
c0de125d80f6006a3559c86fbbfa6a759e4e2ebb
fix(ext/node): make tty ReadStream/WriteStream callable without `new` (#32537) ## Summary - Convert `ReadStream` (in `tty.js`) and `WriteStream` (in `internal/tty.js`) from ES classes to function-style constructors so they can be called without `new`, matching Node.js behavior - Uses `FunctionPrototypeCall` and `Obje...
[ { "path": "ext/node/polyfills/internal/tty.js", "patch": "@@ -8,9 +8,11 @@ import { primordials } from \"ext:core/mod.js\";\n const {\n ArrayPrototypeSome,\n Error,\n+ FunctionPrototypeCall,\n ObjectEntries,\n ObjectPrototypeHasOwnProperty,\n ObjectPrototypeIsPrototypeOf,\n+ ObjectSetPrototypeOf...
2026-03-10T11:35:33
ollama/ollama
8711d03df7da3b19f2ffa58957bf7f872a37b0ad
6e8db04716085c86aae07fc9e72fb41fefdaedc4
Report errors on server lookup instead of path lookup failure
[ { "path": "llm/server.go", "patch": "@@ -253,6 +253,12 @@ func NewLlamaServer(model string, adapters, projectors []string, opts api.Option\n \tvar finalErr error\n \tfor i := 0; i < len(servers); i++ {\n \t\tdir := availableServers[servers[i]]\n+\t\tif dir == \"\" {\n+\t\t\t// Shouldn't happen\n+\t\t\tfinal...
2024-04-22T23:22:05
vuejs/vue
bacb911f7df09ff4868b4c848a6d7778872dff5c
c0d516c283aa1a1c238b6eb8b8e55f64770d27e8
fix(warning): allow symbol as vdom key (#7271)
[ { "path": "src/core/vdom/patch.js", "patch": "@@ -270,7 +270,7 @@ export function createPatchFunction (backend) {\n createElm(children[i], insertedVnodeQueue, vnode.elm, null, true)\n }\n } else if (isPrimitive(vnode.text)) {\n- nodeOps.appendChild(vnode.elm, nodeOps.createTextNode(vn...
2017-12-19T14:35:50
ggml-org/llama.cpp
3d86c6c2b546ae5ce9cdee3ca0ac4a8181e0a073
9963b81f6392da8066958c177db77ad4b4a8f284
model: support GLM4V vision encoder (#18042) * convert ok * no deepstack * less new tensors * cgraph ok * add mrope for text model * faster patch merger * add GGML_ROPE_TYPE_MRNORM * add support for metal * move glm4v do dedicated graph * convert: add norm_embd * clip: add debugging fn * working correctly ...
[ { "path": "convert_hf_to_gguf.py", "patch": "@@ -862,6 +862,14 @@ def set_gguf_parameters(self):\n logger.warning(f\"Unknown RoPE type: {rope_type}\")\n logger.info(f\"gguf: rope scaling type = {rope_gguf_type.name}\")\n \n+ if \"mrope_section\" in self.rope_parameters:\n+...
2025-12-16T10:25:26
huggingface/transformers
3f17410925a0c62f6844f1395270bf0066133cbf
53d2bf6dab1c0ab09ed23889eab43ae92c1ebcd2
Kernel mapping error resolve (#42466) * mapping error resolved with test check * Fix undefined variable 'device' in kernel_config * added test in test_kernels * added test with proper format * added test with proper format once again * Removed mapping_test.py file * reformated with ruff * removed the test
[ { "path": "src/transformers/utils/kernel_config.py", "patch": "@@ -208,6 +208,7 @@ def create_compatible_mapping(self, model, compile=False):\n from kernels import Mode\n \n compatible_mapping = {}\n+ current_device = infer_device(model)\n for layer_name, kernel in self.kernel...
2025-12-02T17:14:29
ollama/ollama
435cc866a3fbabb5029b8a2496631847a871616f
62be2050dd83197864d771fe6891fc47486ee6a1
fix: mixtral graph
[ { "path": "llm/ggml.go", "patch": "@@ -343,7 +343,15 @@ func (llm GGML) GraphSize(context, batch uint64) (partialOffload, fullOffload ui\n \t\t\t4*batch*(embedding+vocab)+embedding*vocab*105/128,\n \t\t)\n \n-\t\tif ffnGateWeight, ok := layers[\"0\"][\"ffn_gate.0.weight\"]; ok {\n+\t\tif ffnGateExpsWeight, ...
2024-04-22T23:57:05
huggingface/transformers
53d2bf6dab1c0ab09ed23889eab43ae92c1ebcd2
5690f24e66a87ee0934659952bacc0d676f57ae4
fix regression (#42569) * fix regression * fix * fix
[ { "path": "src/transformers/integrations/finegrained_fp8.py", "patch": "@@ -316,21 +316,17 @@ def __init__(\n ):\n super().__init__(in_features, out_features)\n \n- # If block size, is not passed, it means that we are doing per-tensor quantization\n- if block_size is not None:\n- ...
2025-12-02T17:06:14
vuejs/vue
c0d516c283aa1a1c238b6eb8b8e55f64770d27e8
01c07503bf6af902dde06fafa8a0008ee3e303aa
feat(types): extract VueConfiguration type for easy expansion (#7273) (#7274) fix #7273
[ { "path": "types/vue.d.ts", "patch": "@@ -64,6 +64,18 @@ export interface Vue {\n export type CombinedVueInstance<Instance extends Vue, Data, Methods, Computed, Props> = Data & Methods & Computed & Props & Instance;\n export type ExtendedVue<Instance extends Vue, Data, Methods, Computed, Props> = VueConstr...
2017-12-19T14:34:35
huggingface/transformers
5690f24e66a87ee0934659952bacc0d676f57ae4
0ba8f001fe826379cb5cf3ce1079b3523bee0381
Fix `parse_response` after tokenizer refactor (#42300) * Remove references to parse_response in the docs * Re-add parse_response (with appropriate warnings) * future annotations * Guard the text generation pipeline correctly
[ { "path": "docs/source/en/_toctree.yml", "patch": "@@ -88,8 +88,6 @@\n title: Tool use\n - local: chat_templating_writing\n title: Writing a chat template\n- - local: chat_response_parsing\n- title: Response parsing\n title: Chat with models\n - sections:\n - local: serving...
2025-12-02T16:46:16
vuejs/vue
d544d052a9c5ec113c253895211296120d58b6ab
b8d33ecd9ab8b7a46d8558b4e2caf506235cd165
fix(weex): append as tree by default for recycle-list and cell-slot (#7216)
[ { "path": "src/platforms/weex/compiler/modules/append.js", "patch": "@@ -1,7 +1,13 @@\n /* @flow */\n \n+import { makeMap } from 'shared/util'\n+\n+// The \"unitary tag\" means that the tag node and its children\n+// must be sent to the native together.\n+const isUnitaryTag = makeMap('cell,header,cell-slot,...
2017-12-11T14:48:36
ollama/ollama
9c0db4cc83d488a9a51d262fac34aea1f651d347
6f18297b3a8711da2a538a0cd670e7dae48307a6
Update gen_windows.ps1 Fixed improper env references
[ { "path": "llm/generate/gen_windows.ps1", "patch": "@@ -243,9 +243,9 @@ if ($null -ne $script:CUDA_LIB_DIR) {\n init_vars\n $script:buildDir=\"../build/windows/${script:ARCH}/cuda$script:CUDA_VARIANT\"\n $script:cmakeDefs += @(\"-A\", \"x64\", \"-DLLAMA_CUDA=ON\", \"-DLLAMA_AVX=on\", \"-DLLAMA_A...
2024-04-21T20:13:41
denoland/deno
c0de125d80f6006a3559c86fbbfa6a759e4e2ebb
a33c9748401e2261c46211c247f819564c4218d3
fix(ext/node): add missing http2 header constants (#32611) Adds missing HTTP2_HEADER_ACCEPT_ENCODING and related constants that grpc-js depends on. Without these, grpc-js server crashes with: ``` TypeError: Cannot read properties of undefined (reading 'toLowerCase') at normalizeKey (metadata.js:38:16) at Meta...
[ { "path": "ext/node/ops/http2/types.rs", "patch": "@@ -152,6 +152,10 @@ pub struct Http2Constants {\n http2_header_transfer_encoding: &'static str,\n http2_header_keep_alive: &'static str,\n http2_header_proxy_connection: &'static str,\n+ http2_header_accept: &'static str,\n+ http2_header_accept_enc...
2026-03-10T11:07:10
huggingface/transformers
0ba8f001fe826379cb5cf3ce1079b3523bee0381
5efd0d4aa5ba4455f229b10667718cc623ac56af
Add backward compatibility for methods which have been moved to `RotaryEmbeddingConfigMixin` (#42517) * Add backward compatibility for methods which have been moved to `RotaryEmbeddingConfigMixin` Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> * They're not actually no-ops Signed-off-by: Ha...
[ { "path": "src/transformers/modeling_rope_utils.py", "patch": "@@ -13,6 +13,7 @@\n # limitations under the License.\n \n import math\n+import warnings\n from functools import wraps\n from typing import TYPE_CHECKING, Optional, TypedDict\n \n@@ -656,7 +657,7 @@ def standardize_rope_params(self):\n # ...
2025-12-02T16:28:37
vuejs/vue
5c2ce0017ff8929e70ce9f701b91d950fb351adb
801f793625273b39fd3f25abbaa04508d6651563
feat(weex): WIP fix flow + handle errors in recycle-list template render
[ { "path": "flow/compiler.js", "patch": "@@ -21,9 +21,6 @@ declare type CompilerOptions = {\n shouldDecodeNewlinesForHref?: boolean;\n optimize?: boolean;\n \n- // support <recycle-list> in weex\n- recyclable?: boolean;\n-\n // for ssr optimization compiler\n scopeId?: string;\n \n@@ -37,7 +34,6 @@...
2017-11-29T15:35:20
ollama/ollama
62be2050dd83197864d771fe6891fc47486ee6a1
56f8aa6912870123930ae0e15e6245ebbb557bd3
chore: use errors.New to replace fmt.Errorf will much better (#3789)
[ { "path": "api/types.go", "patch": "@@ -2,6 +2,7 @@ package api\n \n import (\n \t\"encoding/json\"\n+\t\"errors\"\n \t\"fmt\"\n \t\"math\"\n \t\"os\"\n@@ -307,7 +308,7 @@ func (m *Metrics) Summary() {\n \t}\n }\n \n-var ErrInvalidOpts = fmt.Errorf(\"invalid options\")\n+var ErrInvalidOpts = errors.New(\"in...
2024-04-21T02:11:06
ggml-org/llama.cpp
db81d5ec4b0a9cb19e98c4533731c9554eb025db
c05aa69f32ee87b2019a1b39979d89302c2d6b7b
model-conversion : use CONVERTED_EMBEDDING_MODEL for embedding_verify_logits (#18079) This commit updates the embedding model verification script to use the CONVERTED_EMBEDDING_MODEL environment variable instead of using the EMBEDDING_MODEL_PATH (the original embedding model path) as the basis for the converted model ...
[ { "path": "examples/model-conversion/scripts/embedding/compare-embeddings-logits.sh", "patch": "@@ -34,8 +34,11 @@ done\n MODEL_PATH=\"${MODEL_PATH:-\"$EMBEDDING_MODEL_PATH\"}\"\n MODEL_NAME=\"${MODEL_NAME:-$(basename \"$MODEL_PATH\")}\"\n \n+CONVERTED_MODEL_PATH=\"${CONVERTED_EMBEDDING_PATH:-\"$CONVERTED_E...
2025-12-16T10:17:20
denoland/deno
0eddc94dc9ecc27f4f35da32fe847673a05562c4
913548b879a7530d7992d8722954ac5060194d28
test: ignore some Node compat tests (#32556) ## Summary - Adds an `ignore` field to the node_compat test config schema (`config.jsonc`) so tests can be skipped on all platforms with a single flag instead of setting `darwin: false, linux: false, windows: false` separately - `reason` is required when `ignore: true` is ...
[ { "path": "tests/node_compat/config.jsonc", "patch": "@@ -4,11 +4,17 @@\n \"abort/test-addon-register-signal-handler.js\": {},\n \"abort/test-addon-uv-handle-leak.js\": {},\n \"abort/test-zlib-invalid-internals-usage.js\": {\n- \"darwin\": false,\n- \"linux\": false,\n- \"windows\...
2026-03-10T10:38:04
vuejs/vue
c1743a9dd7eaffa67f9a9c21b6940f167ba7590d
0ee81b24b5146bca315503e2f5aa3b01832735f1
test(weex): use done.fail for reporting errors
[ { "path": "test/weex/cases/cases.spec.js", "patch": "@@ -21,10 +21,7 @@ function createRenderTestCase (name) {\n expect(getRoot(instance)).toEqual(target)\n done()\n }, 50)\n- }).catch(err => {\n- expect(err).toBe(null)\n- done()\n- })\n+ }).catch(done.fail)\n }\n ...
2017-11-28T19:50:41
huggingface/transformers
5efd0d4aa5ba4455f229b10667718cc623ac56af
4ec83fe9f652fae57af0ce03a74da165eb7c38e9
Fix eetq quanto quant methods (#42557) * fix * eetq dep removed * maybe ? * Fix ! * eveyrthing is passing ! * Apply style fixes * move to nn.paramters * grad false * fix * fix --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
[ { "path": "src/transformers/integrations/eetq.py", "patch": "@@ -12,11 +12,13 @@\n # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n # See the License for the specific language governing permissions and\n # limitations under the License.\n-from ..utils import is_accelerate_availab...
2025-12-02T15:57:20
denoland/deno
c28573335523cb26aaa2da09bb8a2a6366e50b77
20e1ef892e7605469a239c269705f417e42bd5d8
fix(ext/node): napi_set_instance_data on exit (#32509) Closes https://github.com/denoland/deno/issues/30529 --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
[ { "path": "ext/napi/lib.rs", "patch": "@@ -306,8 +306,13 @@ impl<T> PendingNapiAsyncWork for T where T: FnOnce() + Send + 'static {}\n pub struct NapiState {\n // Thread safe functions.\n pub env_cleanup_hooks: Rc<RefCell<Vec<(napi_cleanup_hook, *mut c_void)>>>,\n+ pub env_shared_ptrs: Vec<*mut EnvShar...
2026-03-10T08:55:08
ggml-org/llama.cpp
c05aa69f32ee87b2019a1b39979d89302c2d6b7b
279cef27c2b297476ecde6e5df729c057691f1de
common : add nemotron 3 parsing (#18077) * common : expose json-schema functionality to extract type info * common : fix peg parser negation during needs_more_input * common : add some defensive measures in constructed peg parser * common : add nemotron nano 3 support * common : add nemotron nano 3 tests * remove...
[ { "path": "common/chat-peg-parser.cpp", "patch": "@@ -4,9 +4,14 @@\n \n using json = nlohmann::json;\n \n-static std::string_view trim_trailing_space(std::string_view sv) {\n+static std::string_view trim_trailing_space(std::string_view sv, int max = -1) {\n+ int count = 0;\n while (!sv.empty() && std...
2025-12-16T10:05:23
huggingface/transformers
bc7a268fed343ab22446ec86115cf2727b38a5eb
29e8522b85d9bf4287f63611aaa1a7db1fdc6489
Fix fp8 + some enhancement (#42455) * Fix fp8 + some enhancement * style * Add coauthor Co-authored-by: Yang Kai <kai.yang@intel.com> * fix * style * fix tests * style * assertin * style * fix * fix * Apply suggestions from code review Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.git...
[ { "path": "src/transformers/core_model_loading.py", "patch": "@@ -826,6 +826,7 @@ def convert_and_load_state_dict_in_model(\n if hf_quantizer and hf_quantizer.pre_quantized and original_key != renamed_key:\n # if the key was renamed as it is not available in the state dict otherw...
2025-12-02T14:42:01
vuejs/vue
0ee81b24b5146bca315503e2f5aa3b01832735f1
3b32652aa68a06d881e3149bb21ac8711887e7f6
fix(weex): update recycle-list v-for transform
[ { "path": "src/compiler/parser/index.js", "patch": "@@ -23,7 +23,7 @@ export const onRE = /^@|^v-on:/\n export const dirRE = /^v-|^@|^:/\n export const forAliasRE = /(.*?)\\s+(?:in|of)\\s+(.*)/\n export const forIteratorRE = /,([^,\\}\\]]*)(?:,([^,\\}\\]]*))?$/\n-const stripParensRE = /^\\(|\\)$/g\n+export ...
2017-11-27T22:48:46
ggml-org/llama.cpp
c560316440925e03d3ba006eb5710f6678802ce8
d6742125c317b7daafec038ab54a7c2fb1e2beaf
graph : reuse SSM graphs (#16490) * graph : reuse hybrid graphs * graph : reuse recurrent graphs * graph : fix reuse check for recurrent inputs * memory : move the recurrent state into the memory context * Revert "memory : move the recurrent state into the memory context" This reverts commit 00f115fe810815d4a22a6...
[ { "path": "src/llama-graph.cpp", "patch": "@@ -254,6 +254,24 @@ void llm_graph_input_rs::set_input(const llama_ubatch * ubatch) {\n }\n }\n \n+bool llm_graph_input_rs::can_reuse(const llm_graph_params & params) {\n+ const auto * mctx = static_cast<const llama_memory_recurrent_context *>(params.mctx);...
2025-12-16T07:36:21
huggingface/transformers
29e8522b85d9bf4287f63611aaa1a7db1fdc6489
64d8cf4fd957d87d6d1da8ee46badc502b728ba8
[Fix] dots1 expert bias routing (#41663)
[ { "path": "src/transformers/models/dots1/modeling_dots1.py", "patch": "@@ -369,9 +369,11 @@ def __init__(self, config):\n \n def route_tokens_to_experts(self, router_logits):\n router_logits = router_logits.sigmoid() # main diff with deepseekv3\n- router_logits = router_logits + self.gat...
2025-12-02T14:26:45
denoland/deno
20e1ef892e7605469a239c269705f417e42bd5d8
8a3ed27eb1fea2629b67cb2a7a7ef15bd58e3710
fix(ext/node): rewrite http2 implementation (#32418) Rewrite Node.js http2 module using nghttp2 with libuv compatibility layer. Co-authored-by: Divy Srivastava <me@littledivy.com> Co-authored-by: kajukitli <kajukitli@users.noreply.github.com> Co-authored-by: Nathan Whitaker <nathan@deno.com>
[ { "path": "Cargo.lock", "patch": "@@ -595,6 +595,26 @@ dependencies = [\n \"syn 2.0.117\",\n ]\n \n+[[package]]\n+name = \"bindgen\"\n+version = \"0.71.1\"\n+source = \"registry+https://github.com/rust-lang/crates.io-index\"\n+checksum = \"5f58bf3d7db68cfbac37cfc485a8d711e87e064c3d0fe0435b92f7a407f9d6b3\"\...
2026-03-10T08:45:38
vuejs/vue
2cb8ea3fee741807a15bf8f3049ab062a7d9508c
ac999573ea6e4be3a421cdff79d66a5274ec58eb
feat(weex): support compiling `v-on` in the weex native directive (#6892) * refactor(compiler): move postTransforms to after children are processed * feat(weex): recycle-list support WIP * refactor: fix types * feat(weex): split text into separate module * feat($compiler): supports compiling v-bind to the ...
[ { "path": "flow/compiler.js", "patch": "@@ -56,6 +56,7 @@ declare type ASTIfConditions = Array<ASTIfCondition>;\n \n declare type ASTElementHandler = {\n value: string;\n+ params?: Array<any>;\n modifiers: ?ASTModifiers;\n };\n ", "additions": 1, "deletions": 0, "language": "JavaScript" }...
2017-10-31T00:28:40
ggml-org/llama.cpp
3034836d36bfd4ba1d1cc9173c9bef2c70ffd136
a20979d433e3c8b63a6ba20ed5956ced16f65058
webui: Improve copy to clipboard with text attachments (#17969) * feat: Create copy/paste user message including "pasted text" attachments * chore: update webui build output * chore: update webui static output * fix: UI issues * chore: update webui static output * fix: Decode HTML entities using `DOMParser` * ch...
[ { "path": "tools/server/webui/README.md", "patch": "@@ -619,11 +619,12 @@ flowchart TB\n \n ### Test Types\n \n-| Type | Tool | Location | Command |\n-| ------------- | ------------------ | -------------------------------- | ------------------- |\n-...
2025-12-16T06:38:46
huggingface/transformers
64d8cf4fd957d87d6d1da8ee46badc502b728ba8
bb09a30f5ad57fe837526b60d72981a41f08c0fd
[Ministral 3] Small fix config (#42537) * WIP * WIP
[ { "path": "docs/source/en/model_doc/ministral3.md", "patch": "@@ -48,7 +48,7 @@ model_id = \"mistralai/Ministral-3-3B-Instruct-2512\"\n \n tokenizer = MistralCommonBackend.from_pretrained(model_id)\n model = Mistral3ForConditionalGeneration.from_pretrained(\n- model_id, torch_dtype=torch.bfloat16, device...
2025-12-02T13:40:24
denoland/deno
8a3ed27eb1fea2629b67cb2a7a7ef15bd58e3710
ddbdb6d294eab9f30d6a64b12f2df77b046a2b9b
fix(ext/node): improve node:dns compat (#32536) ## Summary Several improvements to the `node:dns` module compatibility: - Add `ipv6first` as a valid value for `dns.setDefaultResultOrder()` / `dns.promises.setDefaultResultOrder()` - Add `order` option support to `dns.lookup()` and `dns.promises.lookup()`, accepting `...
[ { "path": "ext/net/lib.rs", "patch": "@@ -74,6 +74,7 @@ deno_core::extension!(deno_net,\n ops::op_net_set_broadcast_udp,\n ops::op_net_validate_multicast,\n ops::op_dns_resolve,\n+ ops::op_net_get_system_dns_servers,\n ops::op_set_nodelay,\n ops::op_set_keepalive,\n ops::op_net_li...
2026-03-09T22:41:53
vuejs/vue
08660e876d5fbdbeed7e1c779c902969814af3e6
5254ee31c481ac16cf8f822b0b4df0f7815ffff3
refactor: fix types
[ { "path": "src/platforms/weex/compiler/modules/recycle-list/index.js", "patch": "@@ -33,9 +33,11 @@ function postTransformNode (el: ASTElement) {\n function genText (node) {\n const value = node.type === 3\n ? node.text\n- : node.tokens.length === 1\n- ? node.tokens[0]\n- : node.tokens\n+...
2017-09-06T09:48:01
ollama/ollama
c942e4a07b91dc6b78bb245241ea514b752e3d4d
bd54b08261c15e927234d03e2b1020e528b38afe
Fixed startup sequence to report model loading
[ { "path": "llm/ext_server/server.cpp", "patch": "@@ -2726,7 +2726,7 @@ static json format_detokenized_response(std::string content)\n static void log_server_request(const httplib::Request &req, const httplib::Response &res)\n {\n // skip GH copilot requests when using default port\n- if (req.path == ...
2024-04-17T15:40:32
ggml-org/llama.cpp
a20979d433e3c8b63a6ba20ed5956ced16f65058
2995341730f18deb64faa4538bda113328fd791f
webui: Add setting to always show sidebar on Desktop (#17809) * feat: Add setting to always show Sidebar on Desktop * chore: update webui build output * feat: Add auto-show sidebar setting * fix: Mobile settings dialog UI * chore: update webui build output * feat: UI label update * chore: update webui build outp...
[ { "path": "tools/server/webui/src/lib/components/app/chat/ChatSettings/ChatSettings.svelte", "patch": "@@ -109,6 +109,16 @@\n \t\t\t\t\tkey: 'disableAutoScroll',\n \t\t\t\t\tlabel: 'Disable automatic scroll',\n \t\t\t\t\ttype: 'checkbox'\n+\t\t\t\t},\n+\t\t\t\t{\n+\t\t\t\t\tkey: 'alwaysShowSidebarOnDesktop'...
2025-12-16T06:31:37
huggingface/transformers
bb09a30f5ad57fe837526b60d72981a41f08c0fd
ac0769cd8ca186d4dea6a04cd6e87c38c14a34b1
[Quantization] fix dequant when block size is none & static quantization (#42545) * fix * style
[ { "path": "src/transformers/integrations/finegrained_fp8.py", "patch": "@@ -686,7 +686,7 @@ def convert(\n missing_keys=None,\n **kwargs,\n ) -> dict[str, torch.Tensor]:\n- if len(input_dict) != 2:\n+ if len(input_dict) < 2:\n # in case of no scales, the weights...
2025-12-02T13:37:39
denoland/deno
ddbdb6d294eab9f30d6a64b12f2df77b046a2b9b
d77f623c4e8764b1c74ed6248c11fb9737e1f17c
fix(ext/node): emit correct error codes for unsupported PQC key types (#32533) ## Summary - When `createPublicKey` or `createPrivateKey` encounters an unsupported PQC key type (ML-KEM, ML-DSA, SLH-DSA), emit Node-compatible error codes (`ERR_OSSL_EVP_DECODE_ERROR` / `ERR_OSSL_UNSUPPORTED`) instead of a plain TypeError...
[ { "path": "ext/node_crypto/keys.rs", "patch": "@@ -611,6 +611,12 @@ pub enum EdRawError {\n UnsupportedCurve,\n }\n \n+#[derive(Debug, thiserror::Error, deno_error::JsError)]\n+#[class(generic)]\n+#[error(\"unsupported\")]\n+#[property(\"code\" = \"ERR_OSSL_UNSUPPORTED\")]\n+pub struct UnsupportedPrivateK...
2026-03-09T22:39:20
vuejs/vue
3f0c628e2c0fe6bfaecc521c96c6cc12ff24c7c4
90ed48224e0ae281a2579b997e4bd5a150b80413
fix: avoid unnecessary lowercase coersion in component name validation close #7237
[ { "path": "src/core/util/options.js", "patch": "@@ -260,8 +260,7 @@ export function validateComponentName (name: string) {\n 'and must start with a letter.'\n )\n }\n- const lower = name.toLowerCase()\n- if (isBuiltInTag(lower) || config.isReservedTag(lower)) {\n+ if (isBuiltInTag(name) || co...
2017-12-14T16:52:37
ollama/ollama
e74163af4cdfeba30d8e51777644d0114506cda1
fb9580df85c562295d919b6c2632117d3d8cea89
fix padding to only return padding
[ { "path": "llm/gguf.go", "patch": "@@ -248,13 +248,17 @@ func (llm *gguf) Decode(rs io.ReadSeeker) error {\n \t}\n \n \tpadding := llm.padding(offset, int64(alignment))\n-\tif _, err := rs.Seek(padding-offset, io.SeekCurrent); err != nil {\n+\tif _, err := rs.Seek(padding, io.SeekCurrent); err != nil {\n \t...
2024-04-16T00:31:11
ggml-org/llama.cpp
40d9c394f4523bd5f214e6ea536655c1e9a83672
d6a1e18c651a46109cbf2ad3b299581f0651128f
Webui: Disable attachment button and model selector button when prompt textbox is disabled. (#17925) * Pass disabled state to the file attachments button and the model selector button. * Update index.html.gz * Fix model info card in non-router mode. * Update index.html.gz
[ { "path": "tools/server/webui/src/lib/components/app/chat/ChatForm/ChatFormActions/ChatFormActionFileAttachments.svelte", "patch": "@@ -35,7 +35,7 @@\n \n <div class=\"flex items-center gap-1 {className}\">\n \t<DropdownMenu.Root>\n-\t\t<DropdownMenu.Trigger name=\"Attach files\">\n+\t\t<DropdownMenu.Trigge...
2025-12-16T06:15:49
huggingface/transformers
ac0769cd8ca186d4dea6a04cd6e87c38c14a34b1
57eeb9caa2d661b90b4fee6c7f700ede341677fa
Fix failing test in Glm4vMoeIntegrationTest (#42488) * Fix missing model attribute in Glm4vMoeIntegrationTest * Removed extra condition. --------- Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>
[ { "path": "tests/models/glm4v_moe/test_modeling_glm4v_moe.py", "patch": "@@ -294,7 +294,9 @@ def test_inputs_embeds_matches_input_ids(self):\n @require_torch\n @slow\n class Glm4vMoeIntegrationTest(unittest.TestCase):\n- model = None\n+ @classmethod\n+ def setUpClass(cls):\n+ cls.model = Non...
2025-12-02T12:55:28
vuejs/vue
14f7015578db8a740450b8f1f45e74b1604ff173
531371b818b0e31a989a06df43789728f23dc4e8
test: fix ssr test due to webpack asset emit order change
[ { "path": "test/ssr/ssr-template.spec.js", "patch": "@@ -227,8 +227,8 @@ describe('SSR: template option', () => {\n `<link rel=\"preload\" href=\"/0.js\" as=\"script\">` +\n `<link rel=\"preload\" href=\"/test.css\" as=\"style\">` +\n // images and fonts are only preloaded when explicitly ...
2017-12-13T00:16:17
denoland/deno
d77f623c4e8764b1c74ed6248c11fb9737e1f17c
d7e449d85de2b1414bd7ce110d890862104e3fb2
fix(console): prevent multi-line object values in console.table (#32604) ## Summary - Objects in `console.table` could break the table layout when their `Deno.inspect` output exceeded the default `breakLength` (72 chars), causing some rows to wrap across multiple lines - Fix: set `breakLength: Infinity` in the inspec...
[ { "path": "ext/web/01_console.js", "patch": "@@ -3771,6 +3771,7 @@ class Console {\n ...getConsoleInspectOptions(noColorStdout()),\n depth: 1,\n compact: true,\n+ breakLength: Infinity,\n });\n const toTable = (header, body) => this.log(cliTable(header, body));\n ", ...
2026-03-09T22:28:20
huggingface/transformers
57eeb9caa2d661b90b4fee6c7f700ede341677fa
a649767573430f2f9dbc81ac3a801f8546740109
Fix Qwen-VL family with prompt tuning (#42508) * fix all qwen models with promp tuning * forgot to rename * fix style * fallback better when no cache position * just why?
[ { "path": "src/transformers/models/qwen2_5_omni/modeling_qwen2_5_omni.py", "patch": "@@ -1958,11 +1958,8 @@ def forward(\n audio_feature_lengths = None\n \n if attention_mask is not None and position_ids is None:\n- if (\n- cache_position is None\n- ...
2025-12-02T11:05:37
ggml-org/llama.cpp
c45f89d5516221e51eef93e467e3bec6eb811994
9d52f17ae33e8df958e20f3f1b13bfec53ab5a1d
ggml-hexagon: mm for mtmd (#17894) * feat: add run_mtmd script for hexagon * fix: fix issue in fp16xfp32 mm * fix: remove opt_experiment for fp16xfp32 mm * fix: ggml-hexagon: matmul fp16xfp32 support non-contigious src0 * fix: fix syntax check for run-mtmd.sh for cli
[ { "path": "ggml/src/ggml-hexagon/ggml-hexagon.cpp", "patch": "@@ -1976,9 +1976,6 @@ static bool ggml_hexagon_supported_mul_mat(const struct ggml_hexagon_session * s\n break;\n \n case GGML_TYPE_F16:\n- if (!opt_experimental) {\n- return false;\n- }\n ...
2025-12-15T18:53:56
denoland/deno
e76e5fa173945757aab4674573772b8793fa20a8
881f8c0e0f0f3431185247a0e28fa41a0c079cfb
fix(ext/node): multiple readline improvements (#32538) ## Summary Several fixes and feature additions to the Node.js readline polyfill: - **Fix tab completion column formatting** — `Math.max.apply(completionsWidth)` was missing the array argument, causing all completions to display one-per-line instead of in columns...
[ { "path": "ext/node/polyfills/_readline.mjs", "patch": "@@ -133,12 +133,13 @@ Interface.prototype.question = function question(query, options, cb) {\n };\n options.signal.addEventListener(\"abort\", onAbort, { once: true });\n const cleanup = () => {\n- options.signal.removeEventListener(on...
2026-03-09T22:07:20
vuejs/vue
8936b8d9c147441555fcfd4ac748d817ba5ff60e
86e4d7566e685f1a98ffab1777fa7d07603c994a
fix(ssr): properly handle errors in async component fix #6778
[ { "path": "src/server/create-renderer.js", "patch": "@@ -78,11 +78,15 @@ export function createRenderer ({\n return false\n }, cb)\n try {\n- render(component, write, context, () => {\n+ render(component, write, context, err => {\n if (template) {\n re...
2017-12-12T23:12:02
huggingface/transformers
a649767573430f2f9dbc81ac3a801f8546740109
0fa49db1205e0a2745161ccac46184e7e46b6e2b
Add FastVLM (#41112) * Added an initial conversion script * Added a modular where FastVLM is different from LlaVA * Improved the conversion script * Adjusted the conversion script * Removed redundant labels from FastViT & improved the template * Added docs and changed default config * Fix default config * Fix d...
[ { "path": "docs/source/en/_toctree.yml", "patch": "@@ -1030,6 +1030,8 @@\n title: Emu3\n - local: model_doc/evolla\n title: Evolla\n+ - local: model_doc/fast_vlm\n+ title: FastVLM\n - local: model_doc/flava\n title: FLAVA\n - local: model_doc/florence2",...
2025-12-02T10:55:35
denoland/deno
881f8c0e0f0f3431185247a0e28fa41a0c079cfb
d0c9173d9e0e92a40b74e03096d2fdec0bd00193
fix(ext/node): use max salt length as default for RSA-PSS signing (#32534) ## Summary - When no `saltLength` is specified for RSA-PSS signing, Node.js defaults to `RSA_PSS_SALTLEN_MAX_SIGN` (max possible salt = `key_bytes - hash_len - 2`) - Deno was defaulting to the digest length instead (via `rsa::pss::Pss::new::<D>...
[ { "path": "ext/node_crypto/sign.rs", "patch": "@@ -8,6 +8,7 @@ use elliptic_curve::generic_array::ArrayLength;\n use rand::rngs::OsRng;\n use rsa::signature::hazmat::PrehashSigner as _;\n use rsa::signature::hazmat::PrehashVerifier as _;\n+use rsa::traits::PublicKeyParts as _;\n use rsa::traits::SignatureSc...
2026-03-09T22:06:55
ggml-org/llama.cpp
9d52f17ae33e8df958e20f3f1b13bfec53ab5a1d
4529c660c89a8675686cbe2bd1839e85f223f3fc
model : add KORMo model (#18032) * vocab: add KORMo Tokenizer * model: add KORMoForCausalLM * vocab: change pretokenizer to qwen2 * lint: fix unintended line removal * model: make qwen2 bias tensor optional * model: use qwen2 architecture for KORMo
[ { "path": "convert_hf_to_gguf.py", "patch": "@@ -1203,6 +1203,9 @@ def get_vocab_base_pre(self, tokenizer) -> str:\n if chkhsh == \"f4f37b6c8eb9ea29b3eac6bb8c8487c5ab7885f8d8022e67edc1c68ce8403e95\":\n # ref: https://huggingface.co/MiniMaxAI/MiniMax-M2\n res = \"minimax-m2\"\...
2025-12-15T17:51:43
ollama/ollama
969238b19e9f6c564fe419424ab60c20ff779bb7
0efb7931c7bc5c140c019238aa88b9a41d382e61
fix padding in decode TODO: update padding() to _only_ returning the padding
[ { "path": "llm/gguf.go", "patch": "@@ -243,7 +243,7 @@ func (llm *gguf) Decode(rs io.ReadSeeker) error {\n \t}\n \n \tpadding := llm.padding(offset, int64(alignment))\n-\tif _, err := rs.Seek(padding, io.SeekCurrent); err != nil {\n+\tif _, err := rs.Seek(padding-offset, io.SeekCurrent); err != nil {\n \t\t...
2024-04-16T00:26:59
vuejs/vue
73a89bf9e53c0f7f00f193e1b1bb195a71ab761f
023f171f58f7f1b36f0b3e69fc6d330366bfdf43
fix(ssr): fix double escaping of ssrNode attribute values (#7224) This fixes a double escaping of attribute values in the SSR optimizing compiler by unescaping the value in `genAttrSegment` because literal attribute values get escaped early during `processAttrs` before it is known, if this attribute will be optimiz...
[ { "path": "src/server/optimizing-compiler/modules.js", "patch": "@@ -77,7 +77,7 @@ function genAttrSegment (name: string, value: string): StringSegment {\n ? ` ${name}=\"${name}\"`\n : value === '\"\"'\n ? ` ${name}`\n- : ` ${name}=${value}`\n+ : ` ${name}=\"${JSO...
2017-12-12T22:06:11
denoland/deno
d0c9173d9e0e92a40b74e03096d2fdec0bd00193
0ee75392669d26bc1e051ecd5bdc90b1997ad963
fix(ext/node): initialize debuglog testEnabled with safe default (#32598) ## Summary - Initialize `testEnabled` and `debugImpls` in `debuglog.ts` with safe defaults (`() => false` and empty object) instead of leaving them deliberately uninitialized - Fixes bootstrap crash `TypeError: testEnabled is not a function` th...
[ { "path": "ext/node/polyfills/internal/util/debuglog.ts", "patch": "@@ -4,10 +4,15 @@\n // TODO(petamoriken): enable prefer-primordials for node polyfills\n // deno-lint-ignore-file prefer-primordials\n \n-// `debugImpls` and `testEnabled` are deliberately not initialized so any call\n-// to `debuglog()` be...
2026-03-09T20:45:41
huggingface/transformers
0fa49db1205e0a2745161ccac46184e7e46b6e2b
bf3f0ae70d0e902efab4b8517fce88f6697636ce
Fix ernie moe (#42535) * fix * style
[ { "path": "src/transformers/models/ernie4_5_moe/modeling_ernie4_5_moe.py", "patch": "@@ -318,51 +318,41 @@ def forward(self, hidden_states):\n \n \n class Ernie4_5_MoeExperts(nn.Module):\n+ \"\"\"Collection of expert weights stored as 3D tensors.\"\"\"\n+\n def __init__(self, config):\n super...
2025-12-01T18:04:32
ggml-org/llama.cpp
4529c660c89a8675686cbe2bd1839e85f223f3fc
0f4f35e7be14d5ef8157aa710b66388e77e99bec
kv-cache: Fix state restore fragmented cache (#17982) * kv-cache : fix state restore with fragmented cache (#17527) Change find_slot to allow non-contiguous allocation during state restore. Fixes 'failed to find available cells in kv cache' error when restoring state to fragmented cache. * tests : update logic * cl...
[ { "path": "src/llama-kv-cache.cpp", "patch": "@@ -1561,9 +1561,11 @@ void llama_kv_cache::state_read(llama_io_read_i & io, llama_seq_id seq_id, llama\n \n const uint32_t strm = seq_id == -1 ? s : seq_to_stream[seq_id];\n \n+ slot_info sinfo;\n+\n bool res = true;\n- res = res &...
2025-12-15T17:28:35
vuejs/vue
023f171f58f7f1b36f0b3e69fc6d330366bfdf43
644274cbd34e14e74e8931fa979b22dc2db04895
fix(core): warn duplicate keys in all cases (#7200) close #7199
[ { "path": "src/core/vdom/patch.js", "patch": "@@ -263,6 +263,9 @@ export function createPatchFunction (backend) {\n \n function createChildren (vnode, children, insertedVnodeQueue) {\n if (Array.isArray(children)) {\n+ if (process.env.NODE_ENV !== 'production') {\n+ checkDuplicateKeys(chil...
2017-12-12T21:38:32