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
rust-lang/rust
9befcfd0f4671675341e770c5bb3bf696c7aa670
da57f1418f0e302e43c757c7f93cad688a895e57
Fix `cross-lang-lto` to not use `path_file_prefix`
[ { "path": "tests/run-make/cross-lang-lto/rmake.rs", "patch": "@@ -3,8 +3,6 @@\n // -Clinker-plugin-lto.\n // See https://github.com/rust-lang/rust/pull/50000\n \n-#![feature(path_file_prefix)]\n-\n use std::path::PathBuf;\n \n use run_make_support::{\n@@ -92,10 +90,17 @@ fn check_bitcode(instructions: LibBu...
2025-02-24T11:37:10
huggingface/transformers
6d4306160ab22c54a677e07f88c7d8808b137d38
35b112d3442e4a9f3873191dcde7d1235b6c72eb
GGUF: Fix llama 3 GGUF (#31358) * Create push-important-models.yml * llama3 support for GGUF * fixup * Update src/transformers/integrations/ggml.py * fix pre-tokenizer * fix * fix * fix * fix * fix * fix * address final comment * handle special tokens + add tests
[ { "path": "src/transformers/integrations/ggml.py", "patch": "@@ -21,7 +21,7 @@\n from array import array\n \n import numpy as np\n-from tokenizers import Tokenizer, decoders\n+from tokenizers import Tokenizer, decoders, normalizers, pre_tokenizers\n from tokenizers.models import BPE\n \n from .. import Adde...
2024-06-20T12:29:58
golang/go
532c199ee56cdbc2cfd12da1c1cfb3359b122c7c
614a8b7c8ad42ff8a9bc363f813af2aae046fd0c
runtime: fix sub-uintptr-sized Windows callback arguments The Windows callback support accepts Go functions with arguments that are uintptr-sized or smaller. However, it doesn't implement smaller arguments correctly. It assumes the Windows arguments layout is equivalent to the Go argument layout. This is often true, b...
[ { "path": "src/runtime/syscall_windows.go", "patch": "@@ -19,9 +19,32 @@ var cbs struct {\n \n // winCallback records information about a registered Go callback.\n type winCallback struct {\n-\tfn *funcval // Go function\n-\targsize uintptr // Callback arguments size (in bytes)\n-\tcdecl bool //...
2020-10-17T22:42:03
ggml-org/llama.cpp
94d026227771ab6497a9900fe85847342f26071b
a93c0ef0fa9ad7b6afed0ac5245d854870b8685e
mtmd: add llama-mtmd-debug binary (#20508) * mtmd: add llama-mtmd-debug binary * adapt * fixes * fix compile error * fix windows compile error * rm legacy clip_debug_encode() * add MTMD_API to fix build
[ { "path": "tools/mtmd/CMakeLists.txt", "patch": "@@ -62,6 +62,10 @@ set_target_properties(mtmd\n PROPERTIES\n PUBLIC_HEADER \"${MTMD_PUBLIC_HEADERS}\")\n \n+set_target_properties(mtmd\n+ PROPERTIES\n+ PRIVATE_HEADER debug/mtmd-debug.h)\n+\n install(TARGETS mtmd LIBRARY PUBLIC_HEADER)\n \n if (...
2026-03-14T14:52:29
nodejs/node
d37789deeb5e951500d7fb499602dc100d3212fb
eeab7bc0688256247c47099a90c67741e6637e42
src: perf_hooks: fix wrong sized delete Depending on the allocator, existing code leaks memory. PR-URL: https://github.com/nodejs/node/pull/16898 Reviewed-By: Franziska Hinkelmann <franziska.hinkelmann@gmail.com> Reviewed-By: Anna Henningsen <anna@addaleax.net> Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By...
[ { "path": "src/node_perf.cc", "patch": "@@ -196,7 +196,9 @@ void PerformanceGCCallback(uv_async_t* handle) {\n \n cleanup:\n delete data;\n- auto closeCB = [](uv_handle_t* handle) { delete handle; };\n+ auto closeCB = [](uv_handle_t* handle) {\n+ delete reinterpret_cast<uv_async_t*>(handle);\n+ };\...
2017-11-09T06:08:13
rust-lang/rust
da57f1418f0e302e43c757c7f93cad688a895e57
bd884d8043b9ebd4e2bbae0e31bc4780d2e26c6a
Fix `issue-107495-archive-permissions` to not rely on `rustc_private`
[ { "path": "tests/run-make/issue-107495-archive-permissions/rmake.rs", "patch": "@@ -1,7 +1,5 @@\n-#![feature(rustc_private)]\n-\n #[cfg(unix)]\n-extern crate libc;\n+use run_make_support::libc;\n \n #[cfg(unix)]\n use std::os::unix::fs::PermissionsExt;", "additions": 1, "deletions": 3, "language...
2025-02-24T11:17:40
huggingface/transformers
35b112d3442e4a9f3873191dcde7d1235b6c72eb
0ed3ffcb4461a244b87781a24e5ebd0a78f98142
Fix a teeny-tiny typo in `tokenization_utils_base.py`'s docstring (#31510) Update tokenization_utils_base.py
[ { "path": "src/transformers/tokenization_utils_base.py", "patch": "@@ -3645,7 +3645,7 @@ def truncate_sequences(\n and `convert_tokens_to_ids` methods.\n num_tokens_to_remove (`int`, *optional*, defaults to 0):\n Number of tokens to remove using the truncation str...
2024-06-20T09:35:52
golang/go
614a8b7c8ad42ff8a9bc363f813af2aae046fd0c
bda37a0b8a4e89318901a68492b79cf6531fa2ff
runtime: tidy Windows callback test This simplifies the systematic test of Windows callbacks with different signatures and prepares it for expanded coverage of function signatures. It now returns a result from the Go function and threads it back through C. This simplifies things, but also previously the code could ha...
[ { "path": "src/runtime/syscall_windows_test.go", "patch": "@@ -9,11 +9,13 @@ import (\n \t\"fmt\"\n \t\"internal/syscall/windows/sysdll\"\n \t\"internal/testenv\"\n+\t\"io\"\n \t\"io/ioutil\"\n \t\"math\"\n \t\"os\"\n \t\"os/exec\"\n \t\"path/filepath\"\n+\t\"reflect\"\n \t\"runtime\"\n \t\"strconv\"\n \t\"...
2020-10-17T02:22:20
ggml-org/llama.cpp
0685848bc64793f6181d82ff4b92b358066cb4af
0024a69b701968bc9ba415b81ef6a11f02d94adf
scripts : remove get-wikitext-103.sh (#20543) It doesn't work and no one seems to use it. $ wget https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-103-raw-v1.zip HTTP request sent, awaiting response... 301 Moved Permanently Location: unspecified ERROR: Redirection (301) without location....
[ { "path": "scripts/get-wikitext-103.sh", "patch": "@@ -1,10 +0,0 @@\n-#!/usr/bin/env bash\n-\n-wget https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-103-raw-v1.zip\n-\n-echo \"Usage:\"\n-echo \"\"\n-echo \" ./llama-perplexity -m model.gguf -f wiki.test.raw [other params]\"\n-echo \"\"\n-\n-e...
2026-03-14T10:22:04
nodejs/node
eeab7bc0688256247c47099a90c67741e6637e42
ab64b6d7992efb9c558719b6e7e63f280d73048b
repl: support top-level await Much of the AST visitor code was ported from Chrome DevTools code written by Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>. PR-URL: https://github.com/nodejs/node/pull/15566 Fixes: https://github.com/nodejs/node/issues/13209 Refs: https://chromium.googlesource.com/chromium/src/+/e8111...
[ { "path": "lib/internal/repl/await.js", "patch": "@@ -0,0 +1,128 @@\n+'use strict';\n+\n+const acorn = require('internal/deps/acorn/dist/acorn');\n+const walk = require('internal/deps/acorn/dist/walk');\n+\n+const noop = () => {};\n+const visitorsWithoutAncestors = {\n+ ClassDeclaration(node, state, c) {\n...
2017-09-23T07:10:47
huggingface/transformers
0ed3ffcb4461a244b87781a24e5ebd0a78f98142
547b5582ec85147492f2485dd8e9cbbeb1016fd8
Add valid columns check in _remove_unused_columns method (#31466) * Add valid columns checking in _remove_unused_columns method https://github.com/huggingface/datasets/issues/6973#issue-2355517362 https://github.com/huggingface/datasets/issues/6535 https://discuss.huggingface.co/t/indexerror-invalid-key-16-is-out...
[ { "path": "src/transformers/trainer.py", "patch": "@@ -810,6 +810,12 @@ def _remove_unused_columns(self, dataset: \"datasets.Dataset\", description: Optio\n )\n \n columns = [k for k in signature_columns if k in dataset.column_names]\n+ if len(columns) == 0:\n+ raise Va...
2024-06-19T12:26:37
rust-lang/rust
cf0752527f1e1b91cff5683b59e49153187d9efe
e5c1d1cb7ee727246a6d453f52b5707e264387d5
Make synthetic RPITIT assoc ty name handling more rigorous. Currently it relies on special treatment of `kw::Empty`, which is really easy to get wrong. This commit makes the special case clearer in the type system by using `Option`. It's a bit clumsy, but the synthetic name handling itself is a bit clumsy; better to m...
[ { "path": "clippy_utils/src/lib.rs", "patch": "@@ -3489,7 +3489,7 @@ fn maybe_get_relative_path(from: &DefPath, to: &DefPath, max_super: usize) -> St\n // a::b::c ::d::sym refers to\n // e::f::sym:: ::\n // result should be super::super::super::super::e::f\n-...
2025-03-04T03:34:18
golang/go
30c18878730434027dbefd343aad74963a1fdc48
404899f6b56800c1d8e0521fc9ce0c856e459d94
runtime,cmd/cgo: simplify C -> Go call path This redesigns the way calls work from C to exported Go functions. It removes several steps from the call path, makes cmd/cgo no longer sensitive to the Go calling convention, and eliminates the use of reflectcall from cgo. In order to avoid generating a large amount of FFI...
[ { "path": "misc/cgo/test/callback.go", "patch": "@@ -181,7 +181,7 @@ func testCallbackCallers(t *testing.T) {\n \tname := []string{\n \t\t\"runtime.cgocallbackg1\",\n \t\t\"runtime.cgocallbackg\",\n-\t\t\"runtime.cgocallback_gofunc\",\n+\t\t\"runtime.cgocallback\",\n \t\t\"runtime.asmcgocall\",\n \t\t\"runt...
2020-10-01T21:22:38
nodejs/node
61567299e79525e82f4942fb1941e61c633f551f
886c5f47e19f72a01e123fa037d35818e502aa84
test: make REPL test pass in coverage mode Make a REPL tab completion test pass in coverage mode by using `Uin` as the common prefix of all `Uint*Array` globals instead of `co` which could be a prefix for `console` and `coverage`, so it doesn't expand the way it's expected to in coverage mode. PR-URL: https://github....
[ { "path": "test/parallel/test-repl-tab-complete.js", "patch": "@@ -519,8 +519,8 @@ const editor = repl.start({\n editorStream.run(['.clear']);\n editorStream.run(['.editor']);\n \n-editor.completer('co', common.mustCall((error, data) => {\n- assert.deepStrictEqual(data, [['con'], 'co']);\n+editor.completer...
2017-11-16T20:18:30
ggml-org/llama.cpp
9789c4ecdc01d571331c14e5197514b53839de4b
77e20cc1076f351f6ba37b545a20b21c6207cbe5
ggml : add OpenVINO backend (#15307) * Update build doc * Add cgraph tensor output name to OV op name * Update openvino build instructions * Add initial NPU support * draft NPU support version 2: prefill + kvcache * NPU support version 2: prefill + kvcache * Change due to ggml cgraph changes, not correct yet * ...
[ { "path": ".devops/openvino.Dockerfile", "patch": "@@ -0,0 +1,138 @@\n+ARG OPENVINO_VERSION_MAJOR=2026.0\n+ARG OPENVINO_VERSION_FULL=2026.0.0.20965.c6d6a13a886\n+ARG UBUNTU_VERSION=24.04\n+\n+# Optional proxy build arguments - empty by default\n+ARG http_proxy=\n+ARG https_proxy=\n+\n+## Build Image\n+FROM ...
2026-03-14T05:56:55
huggingface/transformers
547b5582ec85147492f2485dd8e9cbbeb1016fd8
3b5fa14fb839c6ea19ba7b4ecfbfbbd1c7e15964
Consider inheritance in type checking for tensors (#31378) * Consider inheritance in type checking for tensors Add an additional check to bypass type assertion when both tensors are torch.Tensor instances. * Fix the quality issue
[ { "path": "src/transformers/trainer_pt_utils.py", "patch": "@@ -131,9 +131,10 @@ def nested_concat(tensors, new_tensors, padding_index=-100):\n Concat the `new_tensors` to `tensors` on the first dim and pad them on the second if needed. Works for tensors or\n nested list/tuples/dict of tensors.\n ...
2024-06-19T12:05:20
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
rust-lang/rust
af92a33deef026844c4594cc5eb7484527403425
79439323843d4131a3685e704d14086d17ce79a1
Make synthetic RPITIT assoc ty name handling more rigorous. Currently it relies on special treatment of `kw::Empty`, which is really easy to get wrong. This commit makes the special case clearer in the type system by using `Option`. It's a bit clumsy, but the synthetic name handling itself is a bit clumsy; better to m...
[ { "path": "compiler/rustc_hir/src/def.rs", "patch": "@@ -254,7 +254,7 @@ impl DefKind {\n }\n \n // Some `DefKind`s require a name, some don't. Panics if one is needed but\n- // not provided.\n+ // not provided. (`AssocTy` is an exception, see below.)\n pub fn def_path_data(self, name: Opt...
2025-03-04T03:34:18
ggml-org/llama.cpp
5a32a9b8a5bddf3c6234fd2eb17b0a7315328f93
3b439504ba49d41c75885d34c1353d195e09e028
Fix data race in CUDA's "cpy" kernel (influences GGML's DUP, CONT operations). (#20507) * Fix datarace in CUDA's "cpy" kernel. * Remove extra barrier by using more of shared memory.
[ { "path": "ggml/src/ggml-cuda/cpy.cu", "patch": "@@ -56,7 +56,8 @@ static __global__ void cpy_scalar_transpose(const char * cx, char * cdst, const\n const int tx = blockIdx.y * CUDA_CPY_TILE_DIM_2D + threadIdx.x; // transpose block offset\n const int ty = blockIdx.x * CUDA_CPY_TILE_DIM_2D + threadI...
2026-03-14T05:19:44
golang/go
404899f6b56800c1d8e0521fc9ce0c856e459d94
c80022204e8fc36ec487888d471de27a5ea47e17
cmd/link: preserve alignment for stackmap symbols Stackmap symbols are content-addressable, so it may be dedup'd with another symbol with same content. We want stackmap symbols 4-byte aligned. But if it dedup's with another symbol with larger alignment, preserve that alignment. Fixes #42071. Change-Id: I1616dd2b0c17...
[ { "path": "src/cmd/link/internal/ld/symtab.go", "patch": "@@ -580,8 +580,12 @@ func (ctxt *Link) symtab(pcln *pclntab) []sym.SymKind {\n \t\t\tsymGroupType[s] = sym.SGOFUNC\n \t\t\tldr.SetAttrNotInSymbolTable(s, true)\n \t\t\tldr.SetCarrierSym(s, symgofunc)\n-\t\t\tconst align = 4\n-\t\t\tldr.SetSymAlign(s,...
2020-10-24T17:05:31
huggingface/transformers
3b5fa14fb839c6ea19ba7b4ecfbfbbd1c7e15964
f4d189441da624131e81753b279177f91d109d0d
Fix `wandb` integration with `SetFit` model (#30021) Fix W&B integration with SetFit model Co-authored-by: PEARCE Timothe <timothe_pearce@ext.connect-tech.sncf>
[ { "path": "src/transformers/integrations/integration_utils.py", "patch": "@@ -753,7 +753,7 @@ def setup(self, args, state, model, **kwargs):\n combined_dict = {**args.to_dict()}\n \n if hasattr(model, \"config\") and model.config is not None:\n- model_config = model.co...
2024-06-19T11:23:05
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
rust-lang/rust
e7bea57ce42f6882c8721ca59d0a96e13e1cbbe3
91a0e1604f343730022bc903cbf201d8b0a86a71
Fix a typo in the crashtest output.
[ { "path": "src/tools/compiletest/src/runtest/crashes.rs", "patch": "@@ -15,7 +15,7 @@ impl TestCx<'_> {\n // if a test does not crash, consider it an error\n if proc_res.status.success() || matches!(proc_res.status.code(), Some(1 | 0)) {\n self.fatal(&format!(\n- \...
2025-03-04T04:04:38
ggml-org/llama.cpp
3b439504ba49d41c75885d34c1353d195e09e028
463b6a963c2de376e102d878a50d26802f15833c
opencl: fix l2_norm (#20480)
[ { "path": "ggml/src/ggml-opencl/kernels/l2_norm.cl", "patch": "@@ -63,7 +63,7 @@ kernel void kernel_l2_norm_f32(\n \n barrier(CLK_LOCAL_MEM_FENCE);\n \n- const float scale = 1.0f/sqrt(max(sum[0], eps));\n+ const float scale = 1.0f/max(sqrt(sum[0]), eps);\n \n for (int i00 = get_local_id(0); i0...
2026-03-14T05:18:52
huggingface/transformers
f4d189441da624131e81753b279177f91d109d0d
4144c354e9f54bde29357031fb3e48b23f60b836
Fix typo: pas_token_id (#30894) Fix typo
[ { "path": "src/transformers/generation/configuration_utils.py", "patch": "@@ -518,7 +518,7 @@ def validate(self, is_init=False):\n if self.pad_token_id is not None and self.pad_token_id < 0:\n warnings.warn(\n f\"`pad_token_id` should be positive but got {self.pad_token_i...
2024-06-19T10:23:08
golang/go
c80022204e8fc36ec487888d471de27a5ea47e17
94887410d41eb29ea4f2354e80a6aae4ee602fcf
net: prefer /etc/hosts over DNS when no /etc/nsswitch.conf is present Do not mimic glibc behavior if /etc/nsswitch.conf is missing. This will will likely be missing on musl libc systems and glibc systems will likely always have it, resulting in localhost lookup being done over DNS rather than from /etc/hosts. Do what...
[ { "path": "src/net/conf.go", "patch": "@@ -202,11 +202,6 @@ func (c *conf) hostLookupOrder(r *Resolver, hostname string) (ret hostLookupOrde\n \t\t\t// illumos defaults to \"nis [NOTFOUND=return] files\"\n \t\t\treturn fallbackOrder\n \t\t}\n-\t\tif c.goos == \"linux\" {\n-\t\t\t// glibc says the default is...
2020-10-16T16:23:54
rust-lang/rust
5cf0c95552a8400046cfe00afd0d6e8c3970fa10
b178f22bd8267155f2e330731a53ef3e4eac28bb
Show interpreter backtrace error on Ctrl+C See https://github.com/rust-lang/rust/pull/111769
[ { "path": "src/tools/miri/src/concurrency/thread.rs", "patch": "@@ -1150,7 +1150,7 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {\n loop {\n if CTRL_C_RECEIVED.load(Relaxed) {\n this.machine.handle_abnormal_termination();\n- std::process:...
2025-03-07T09:20:13
ggml-org/llama.cpp
1430c35948db08a35ba07ff3c2709bd50d58f563
f17b3be63fd9dfa3d12ad3555c977a6430661769
common/parser: gracefully handle undetected tool parser, print error message. (#20286)
[ { "path": "common/chat-auto-parser-generator.cpp", "patch": "@@ -3,6 +3,7 @@\n #include \"chat.h\"\n #include \"common.h\"\n #include \"json-schema-to-grammar.h\"\n+#include \"log.h\"\n #include \"nlohmann/json.hpp\"\n \n #include <stdexcept>\n@@ -182,7 +183,10 @@ common_peg_parser analyze_tools::build_pars...
2026-03-13T19:56:10
huggingface/transformers
4144c354e9f54bde29357031fb3e48b23f60b836
cd5f7c17908a66d71c9ca6871c7479f9f1a75f39
auto-detect device when no device is passed to pipeline (#31398) * fix device * Update src/transformers/pipelines/base.py Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com> * bug fix * add warning --------- Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.githu...
[ { "path": "src/transformers/pipelines/base.py", "patch": "@@ -843,6 +843,17 @@ def __init__(\n device = next(iter(hf_device_map.values()))\n else:\n device = -1\n+ if (\n+ is_torch_mlu_available()\n+ or is_torch...
2024-06-19T10:12:39
golang/go
94887410d41eb29ea4f2354e80a6aae4ee602fcf
7930d39f5802fd93a651a51f905efe58d3bc9caf
path/filepath: allow EvalSymlinks to work on UNC share roots on Windows Fixes #42079 Previously, EvalSymlinks returned an error when called with the root of a UNC share (e.g. \\server\share). This was due to Windows's FindFirstFile function not supporting a share root path. To resolve this, now return early from toN...
[ { "path": "src/path/filepath/path_windows_test.go", "patch": "@@ -413,6 +413,9 @@ func TestToNorm(t *testing.T) {\n \t\t{`{{tmp}}\\test`, `.\\foo\\bar`, `foo\\bar`},\n \t\t{`{{tmp}}\\test`, `foo\\..\\foo\\bar`, `foo\\bar`},\n \t\t{`{{tmp}}\\test`, `FOO\\BAR`, `foo\\bar`},\n+\n+\t\t// test UNC paths\n+\t\t{\...
2020-10-20T15:15:23
nodejs/node
667f8c642d35da18493ab7cd397070c7bd0fbf84
22901d81c905873486de676c8ca9d13f11503878
doc: correct the wrong added meta data PR-URL: https://github.com/nodejs/node/pull/17072 Fixes: https://github.com/nodejs/node/issues/17026 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Myles Borins <myles.borins@gmail.com>
[ { "path": "doc/api/async_hooks.md", "patch": "@@ -76,7 +76,7 @@ function promiseResolve(asyncId) { }\n #### `async_hooks.createHook(callbacks)`\n \n <!-- YAML\n-added: v9.0.0\n+added: v8.1.0\n -->\n \n * `callbacks` {Object} The [Hook Callbacks][] to register", "additions": 1, "deletions": 1, "l...
2017-11-16T14:32:56
ggml-org/llama.cpp
f17b3be63fd9dfa3d12ad3555c977a6430661769
d7ba99c4850bd687621f13329490dc28f28f17c9
llama : fix pooling assertion crash in chunked GDN detection path (#20468) * llama : fix pooling assertion crash in chunked GDN detection path The chunked fused Gated Delta Net detection in sched_reserve() calls graph_reserve(16*n_seqs, n_seqs, n_outputs, ...) where n_outputs = n_seqs. This creates a dimension mismat...
[ { "path": "src/llama-context.cpp", "patch": "@@ -512,7 +512,12 @@ void llama_context::sched_reserve() {\n \n if (cparams.fused_gdn_ch) {\n // more than one token in the batch per sequence in order to take the chunked path\n- auto * gf = graph_reserve(16*n_seqs, n_seqs, n_outpu...
2026-03-13T18:53:42
huggingface/transformers
cd5f7c17908a66d71c9ca6871c7479f9f1a75f39
1c1aec2ef1d6822fae3ffbb973b4c941f65f4ddf
Add docs on zeroshot image classification prompt templates (#31343) * Add docs on pipeline templates * Fix example and comments Update usage tips * Update docs/source/en/tasks/zero_shot_image_classification.md Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com> * Update docs/source...
[ { "path": "docs/source/en/model_doc/siglip.md", "patch": "@@ -29,6 +29,7 @@ The abstract from the paper is the following:\n - Usage of SigLIP is similar to [CLIP](clip). The main difference is the training loss, which does not require a global view of all the pairwise similarities of images and texts within...
2024-06-19T10:11:44
golang/go
7930d39f5802fd93a651a51f905efe58d3bc9caf
d03437a7c228c6842ad81b4d6e7dabea913cdc5f
log/syslog: set local to true if network is any of "unix", or "unixgram" Fixes #41960 Change-Id: I0e0f0e11610dd2658a8f6b7e345a4aae2c19c85d GitHub-Last-Rev: 8cac718e4854773ca411116043b4b832e0468f09 GitHub-Pull-Request: golang/go#42135 Reviewed-on: https://go-review.googlesource.com/c/go/+/264297 Reviewed-by: Emmanuel ...
[ { "path": "src/log/syslog/syslog.go", "patch": "@@ -161,7 +161,10 @@ func (w *Writer) connect() (err error) {\n \t\tvar c net.Conn\n \t\tc, err = net.Dial(w.network, w.raddr)\n \t\tif err == nil {\n-\t\t\tw.conn = &netConn{conn: c}\n+\t\t\tw.conn = &netConn{\n+\t\t\t\tconn: c,\n+\t\t\t\tlocal: w.network ==...
2020-10-25T03:42:20
nodejs/node
22901d81c905873486de676c8ca9d13f11503878
b3127cd537a62fd02a639333193dde5741ac3a68
async_hooks: add destroy event for gced AsyncResources In cases where libraries create AsyncResources which may be emitting more events depending on usage, the only way to ensure that destroy is called properly is by calling it when the resource gets garbage collected. Fixes: https://github.com/nodejs/node/issues/161...
[ { "path": "benchmark/async_hooks/gc-tracking.js", "patch": "@@ -0,0 +1,45 @@\n+'use strict';\n+const common = require('../common.js');\n+const { AsyncResource } = require('async_hooks');\n+\n+const bench = common.createBenchmark(main, {\n+ n: [1e6],\n+ method: [\n+ 'trackingEnabled',\n+ 'trackingDis...
2017-11-09T21:57:04
ggml-org/llama.cpp
d7ba99c4850bd687621f13329490dc28f28f17c9
fbaa95bc29b9005586b76ab3d74f485eec05f282
server: reset counter related to kill-switch on client error (#20513) * server: reset kill-switch on client error This avoids triggering a server kill switch. If the client sends a request that exceeds the configured context size, an appropriate HTTP 400 response is provided and no tokens are generated. However sin...
[ { "path": "tools/server/server-context.cpp", "patch": "@@ -1189,6 +1189,9 @@ struct server_context_impl {\n ? SLOT_STATE_WAIT_OTHER // wait for the parent to process prompt\n : SLOT_STATE_STARTED;\n \n+ // reset server kill-switch counter\n+ n_empty_consecutive = 0;\n+\...
2026-03-13T17:58:09
golang/go
05b6118139d880a5bced23da9d07bdb0db8e7084
bc0b198bd75a8eef45d0965531ba6fa127d0e8ec
debug/dwarf: add support for DWARFv5 to (*Data).Ranges Updates the (*Data).Ranges method to work with DWARFv5 which uses the new debug_rnglists section instead of debug_ranges. This does not include supporting DW_FORM_rnglistx. General support for DWARFv5 was added by CL 175138. Change-Id: I01f919a865616a3ff12f5bf6...
[ { "path": "src/debug/dwarf/const.go", "patch": "@@ -461,3 +461,15 @@ const (\n \tutSplitCompile = 0x05\n \tutSplitType = 0x06\n )\n+\n+// Opcodes for DWARFv5 debug_rnglists section.\n+const (\n+\trleEndOfList = 0x0\n+\trleBaseAddressx = 0x1\n+\trleStartxEndx = 0x2\n+\trleStartxLength = 0x3\n+\trleOf...
2020-06-04T14:59:06
ggml-org/llama.cpp
b5e1212063e3a605bd421c0edfee1a4a59731f65
8f974d2392da4e6fa422a67050e90f1471d72966
ggml : fix typo gmml (#20512) Signed-off-by: Adrien Gallouët <angt@huggingface.co>
[ { "path": "ggml/CMakeLists.txt", "patch": "@@ -253,7 +253,7 @@ option(GGML_OPENCL_PROFILING \"ggml: use OpenCL profiling (increas\n option(GGML_OPENCL_EMBED_KERNELS \"ggml: embed kernels\" ON)\n option(GGML_OPENCL_USE_ADRENO_KERNELS \"ggml: use opt...
2026-03-13T13:36:13
nodejs/node
a0b1e2eca684a21e87dd8ae1d5becb4c85e4ddd4
1b093cb93df05bea0c9adfb6a060d2a4542c1ae0
build: prevent echoing of recipes for test target Currenlty the test target will echo additional information that might not be that useful, for example: make doc-only make[1]: Nothing to be done for `doc-only'. make lint Running JS linter... Running C++ linter... Total errors found: 0 make[2]: Nothing to be done for `...
[ { "path": "Makefile", "patch": "@@ -210,11 +210,11 @@ test: all\n \t$(MAKE) cctest\n else\n test: all\n-\t$(MAKE) build-addons\n-\t$(MAKE) build-addons-napi\n-\t$(MAKE) doc-only\n-\t$(MAKE) lint\n-\t$(MAKE) cctest\n+\t$(MAKE) -s build-addons\n+\t$(MAKE) -s build-addons-napi\n+\t$(MAKE) -s doc-only\n+\t$(MAK...
2017-11-14T07:41:10
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
ggml-org/llama.cpp
73c9eb8ceda397b651dbb6661b2935f0283a2b1d
983df142a99b764f5cbe9acb6b9ec862f861d353
metal : fix l2 norm scale (#20493)
[ { "path": "ggml/src/ggml-metal/ggml-metal-device.m", "patch": "@@ -1156,7 +1156,7 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te\n case GGML_OP_RWKV_WKV7:\n return true;\n case GGML_OP_GATED_DELTA_NET:\n- return op->src[2]->ne[0] % ...
2026-03-13T09:43:20
nodejs/node
9c1b18a59f2514e7bccbd07108b5f1f4e6f7c05e
6c76140a7a61d0030af5e869b4c0e0b038460b05
tty: fix 'resize' event regression It's not wholly clear what commit introduced the regression but between v8.4.0 and v8.5.0 the 'resize' event stopped getting emitted when the tty was resized. The SIGWINCH event listener apparently was being installed before the support code for `process.on('SIGWINCH', ...)` was. F...
[ { "path": "lib/internal/bootstrap_node.js", "patch": "@@ -33,6 +33,7 @@\n \n const _process = NativeModule.require('internal/process');\n _process.setupConfig(NativeModule._source);\n+ _process.setupSignalHandlers();\n NativeModule.require('internal/process/warning').setup();\n NativeModu...
2017-10-15T20:40:00
golang/go
bc0b198bd75a8eef45d0965531ba6fa127d0e8ec
94eda9ae37d9f8a34c44c1ad1a030b4475f79faf
runtime: dump the status of lockedg on error The dumpgstatus() will dump current g's status anyway. When lockedg's status is bad, it's more helpful to dump lockedg's status as well than dumping current g's status twice. Change-Id: If5248cb94b9cdcbf4ceea07562237e1d6ee28489 GitHub-Last-Rev: da814c51ff42f56fb28582f088f4...
[ { "path": "src/runtime/proc.go", "patch": "@@ -2302,8 +2302,8 @@ func stoplockedm() {\n \tmPark()\n \tstatus := readgstatus(_g_.m.lockedg.ptr())\n \tif status&^_Gscan != _Grunnable {\n-\t\tprint(\"runtime:stoplockedm: g is not Grunnable or Gscanrunnable\\n\")\n-\t\tdumpgstatus(_g_)\n+\t\tprint(\"runtime:sto...
2020-10-15T01:43:51
ollama/ollama
ba460802c20311ac9c90e6a750df2939a3defa9b
e54a3c7fcd3a66486c3946a7944f3d7ce2daff6c
examples: add more Go examples using the API (#3599) * examples: go-multimodal * examples: add go-pull-progress * examples: add go-chat * fix
[ { "path": "examples/go-chat/main.go", "patch": "@@ -0,0 +1,51 @@\n+package main\n+\n+import (\n+\t\"context\"\n+\t\"fmt\"\n+\t\"log\"\n+\n+\t\"github.com/ollama/ollama/api\"\n+)\n+\n+func main() {\n+\tclient, err := api.ClientFromEnvironment()\n+\tif err != nil {\n+\t\tlog.Fatal(err)\n+\t}\n+\n+\tmessages :...
2024-04-15T22:34:54
ggml-org/llama.cpp
983df142a99b764f5cbe9acb6b9ec862f861d353
57819b8d4b39d893408e51520dff3d47d1ebb757
convert : fix/suppress pyright errors (#20442) * convert : fix/suppress pyright errors This commit fixes the pyright errors that are generated by pyright for convert_hf_to_gguf.py. The motivation for this is that running this locally generates errors that CI does not, and it can be difficult to spot new errors. One ...
[ { "path": "convert_hf_to_gguf.py", "patch": "@@ -2194,6 +2194,8 @@ def set_gguf_parameters(self):\n def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:\n n_head = self.hparams.get(\"n_head\", self.hparams.get(\"num_attention_heads\"))\n ...
2026-03-13T05:00:52
nodejs/node
6c76140a7a61d0030af5e869b4c0e0b038460b05
8203ce8316358c91bce455bb0d309cb3b5289d15
src: fix SetClientCertEngine() nullptr dereference Introduced in commit 6ee985f311d ("tls: implement clientCertEngine option") which was merged November 11. PR-URL: https://github.com/nodejs/node/pull/16965 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Franziska Hinkelmann <franziska.hinkelmann@gmail.com>...
[ { "path": "src/node_crypto.cc", "patch": "@@ -1341,7 +1341,8 @@ void SecureContext::SetClientCertEngine(\n CHECK_EQ(args.Length(), 1);\n CHECK(args[0]->IsString());\n \n- SecureContext* sc = Unwrap<SecureContext>(args.This());\n+ SecureContext* sc;\n+ ASSIGN_OR_RETURN_UNWRAP(&sc, args.Holder());\n \n...
2017-11-12T14:16:36
golang/go
d05c03536573b8c567d3eefaad2aafbf89678015
57af9745bfad2c20ed6842878e373d6c5b79285a
net/http: fix typo in TestTransportReadToEndReusesConn The test sets a Content-Type where it looks like it wants a Content-Length. The test passes because the Content-Length header is automatically added anyway, but fix the typo and set Content-Length as intended. Change-Id: Ic2af778f82c3e9d58e164892f6ac6ef5745f884f ...
[ { "path": "src/net/http/transport_test.go", "patch": "@@ -412,7 +412,7 @@ func TestTransportReadToEndReusesConn(t *testing.T) {\n \t\t\tw.WriteHeader(200)\n \t\t\tw.(Flusher).Flush()\n \t\t} else {\n-\t\t\tw.Header().Set(\"Content-Type\", strconv.Itoa(len(msg)))\n+\t\t\tw.Header().Set(\"Content-Length\", st...
2020-08-05T16:57:11
ollama/ollama
7027f264fbb3292fa4b53623fbaef5792d0d5f80
9bee3b63b1a8299384dea04c10f191b13a2815c3
app: gracefully shut down `ollama serve` on windows (#3641) * app: gracefully shut down `ollama serve` on windows * fix linter errors * bring back `HideWindow` * remove creation flags * restore `windows.CREATE_NEW_PROCESS_GROUP`
[ { "path": "app/lifecycle/server.go", "patch": "@@ -9,7 +9,6 @@ import (\n \t\"os\"\n \t\"os/exec\"\n \t\"path/filepath\"\n-\t\"syscall\"\n \t\"time\"\n \n \t\"github.com/ollama/ollama/api\"\n@@ -87,19 +86,29 @@ func SpawnServer(ctx context.Context, command string) (chan int, error) {\n \t// Re-wire context ...
2024-04-14T22:33:25
huggingface/transformers
077c139f574089cf27b6279dce25a52be9908144
609e662243d42769f8ef62f2ccb177c5257ae26f
[tests] rename `test_config_object` to `test_ds_config_object` (#31403) fix name
[ { "path": "tests/deepspeed/test_deepspeed.py", "patch": "@@ -891,7 +891,7 @@ def test_load_state_dict_from_zero_checkpoint(self, stage, dtype):\n self.assertEqual(b, b1)\n self.check_trainer_state_are_the_same(state, state1)\n \n- def test_config_object(self):\n+ def test_ds_co...
2024-06-19T09:19:15
ggml-org/llama.cpp
128142fe7dd02cc87e178778ae80affcc8836304
6de1bc631d1a27e1f71fef2e6f8fad4ea3f5cf67
test-backend-ops: allow loading tests from file and parsing model operators into file (#19896) * tests: allow loading test-backend-ops tests from json * add error threshold based on op * add error when file cannot be read * add graph operator json extraction tool * add nb parameter for non-contiguous input tensors...
[ { "path": "common/arg.cpp", "patch": "@@ -2671,7 +2671,8 @@ common_params_context common_params_parser_init(common_params & params, llama_ex\n [](common_params & params, const std::string & value) {\n params.out_file = value;\n }\n- ).set_examples({LLAMA_EXAMPLE_IMATRIX, LLAMA...
2026-03-12T12:26:00
nodejs/node
8203ce8316358c91bce455bb0d309cb3b5289d15
0fdd88a374e23e1dd4a05d93afd5eb0c3b080fd5
src: fix bad sizeof expression It was computing the size of the pointer, not the size of the pointed-to object. Introduced in commit 727b2911eca ("src,dns: refactor cares_wrap to avoid global state".) PR-URL: https://github.com/nodejs/node/pull/17014 Reviewed-By: Anna Henningsen <anna@addaleax.net> Reviewed-By: Coli...
[ { "path": "src/cares_wrap.cc", "patch": "@@ -160,7 +160,7 @@ class ChannelWrap : public AsyncWrap {\n }\n inline node_ares_task_list* task_list() { return &task_list_; }\n \n- size_t self_size() const override { return sizeof(this); }\n+ size_t self_size() const override { return sizeof(*this); }\n \n...
2017-11-14T09:18:32
golang/go
57af9745bfad2c20ed6842878e373d6c5b79285a
a36493eb3c7f8c990f5112f70e66270372a5b12e
crypto/x509: deprecate legacy PEM encryption It's unfortunate that we don't implement PKCS#8 encryption (#8860) so we can't recommend an alternative but PEM encryption is so broken that it's worth deprecating outright. Fixes #41949 Fixes #32777 Change-Id: Ieb46444662adec108d0de3550b693a50545c2344 Reviewed-on: https:...
[ { "path": "src/crypto/x509/pem_decrypt.go", "patch": "@@ -95,7 +95,12 @@ func (c rfc1423Algo) deriveKey(password, salt []byte) []byte {\n \treturn out\n }\n \n-// IsEncryptedPEMBlock returns if the PEM block is password encrypted.\n+// IsEncryptedPEMBlock returns whether the PEM block is password encrypted\...
2020-10-22T16:00:20
ollama/ollama
9bee3b63b1a8299384dea04c10f191b13a2815c3
309aef7fee1a5b825532d58ddfed8acde765d50c
types/model: add path helpers (#3619) This commit adds path helpers for working with Names in URL and file paths. The new helpers are ParseNameFromPath, ParseNameFromFilePath, Name.Path, and Name.FilePath. This commit also adds Name.DisplayLongest, and Name.DisplayLong. Also, be it updates a place where string...
[ { "path": "types/model/name.go", "patch": "@@ -7,6 +7,7 @@ import (\n \t\"hash/maphash\"\n \t\"io\"\n \t\"log/slog\"\n+\t\"path/filepath\"\n \t\"slices\"\n \t\"strings\"\n \t\"sync\"\n@@ -54,6 +55,10 @@ const (\n \tPartBuild\n \tPartDigest\n \n+\t// NumParts is the number of parts in a Name. In this list, i...
2024-04-13T19:59:19
huggingface/transformers
7c71b61daef4da1ec5dc41674718d895176af3be
b275a410057b282495422a4dcf5782418aa484e6
Fix autocast incompatibility in RecurrentGemma (#30832)
[ { "path": "src/transformers/models/recurrent_gemma/modeling_recurrent_gemma.py", "patch": "@@ -254,8 +254,8 @@ def _update_cache(self, key_states, value_states, **cache_kwargs):\n k_out = k_out[:, :, indices]\n v_out = v_out[:, :, indices]\n \n- k_out[:, :, cache_position]...
2024-06-19T07:59:34
ggml-org/llama.cpp
0a10c34dc146b19cc256a66b36321d753899a684
deee23863bb75a67471c62b605248d3ff725c9f0
grammar: Fix grammar root symbol check (#19761) * grammar: fix bad check for root symbol, correct error logging * add tests to demonstrate root symbol check failure
[ { "path": "src/llama-grammar.cpp", "patch": "@@ -1160,13 +1160,13 @@ struct llama_grammar * llama_grammar_init_impl(\n // if there is a grammar, parse it\n // rules will be empty (default) if there are parse errors\n if (!parser.parse(grammar_str) || parser.rules.empty()) {\n- fprintf(std...
2026-03-12T11:04:56
golang/go
a36493eb3c7f8c990f5112f70e66270372a5b12e
64dc25b2db5d5be55f3f2fde3daac6c8a2873235
crypto/tls: add no-shared to openssl build instructions This prevents the custom-built version of openssl prefering the system libraries over the ones compiled with the specified (weak crypto) options necessary to generate the updates. This difference can lead to confusing failures when updating the tests. Fixes #318...
[ { "path": "src/crypto/tls/handshake_test.go", "patch": "@@ -86,7 +86,7 @@ func checkOpenSSLVersion() error {\n \tprintln(\"to update the test data.\")\n \tprintln(\"\")\n \tprintln(\"Configure it with:\")\n-\tprintln(\"./Configure enable-weak-ssl-ciphers\")\n+\tprintln(\"./Configure enable-weak-ssl-ciphers ...
2020-09-29T09:34:08
nodejs/node
fdbb6dd042a560fc5d3b3ad682a13ddf506128e6
1132ea743449f63d6b1dd09f82ec3bb6f3f402f0
module: speed up package.json parsing If the package.json does not contain the string '"main"', skip parsing it to JSON. Note that this changes the behavior of the module loader in the presence of package.json files that don't contain legal JSON. Such files used to throw an exception but now they are simply ignored ...
[ { "path": "lib/module.js", "patch": "@@ -120,6 +120,9 @@ function readPackage(requestPath) {\n return false;\n }\n \n+ if (!/\"main\"/.test(json))\n+ return packageMainCache[requestPath] = undefined;\n+\n try {\n var pkg = packageMainCache[requestPath] = JSON.parse(json).main;\n } catch (e...
2017-11-09T12:26:46
huggingface/transformers
b275a410057b282495422a4dcf5782418aa484e6
22b41b3f8a5cdb37e686d18d8d9a24eb98a331ec
[`GPT2`] Add SDPA support (#31172) * `gpt2` sdpa support * fix (at least) one test, style, repo consistency * fix sdpa mask in forward --> fixes generation * test * test2 * test3 * test4 * simplify shapes for attn mask creation and small comments * hub fail test * benchmarks * flash attn ...
[ { "path": "docs/source/en/model_doc/gpt2.md", "patch": "@@ -127,6 +127,64 @@ Below is an expected speedup diagram that compares pure inference time between t\n <img src=\"https://huggingface.co/datasets/EduardoPacheco/documentation-images/resolve/main/gpt2_flash_attention_2_speedup.jpg\">\n </div>\n \n+\n+#...
2024-06-19T07:40:57
ollama/ollama
08655170aa7b2472656b00a2d5b6bc2e9d0365e6
2b341069a7d00fa4c8da65773a5ccf5606cb29e8
types/model: make ParseName variants less confusing (#3617) Also, fix http stripping bug. Also, improve upon docs about fills and masks.
[ { "path": "types/model/name.go", "patch": "@@ -3,6 +3,7 @@ package model\n import (\n \t\"cmp\"\n \t\"errors\"\n+\t\"fmt\"\n \t\"hash/maphash\"\n \t\"io\"\n \t\"log/slog\"\n@@ -25,11 +26,17 @@ var (\n \n // Defaults\n const (\n-\t// DefaultMask is the default mask used by [Name.DisplayShortest].\n-\tDefault...
2024-04-12T20:57:57
ggml-org/llama.cpp
deee23863bb75a67471c62b605248d3ff725c9f0
c3e3f9e5338e8eb3d5b13d20f20d38b53d9d19ce
vulkan: add GATED_DELTA_NET op support (#20334) * vulkan: add GATED_DELTA_NET op support Implements the fused gated delta net recurrence as a Vulkan compute shader with full support for scalar gate, KDA vector gate, GQA broadcast, multi-token sequences, and permuted (non-contiguous) q/k inputs. Specialization constan...
[ { "path": "ggml/src/ggml-vulkan/ggml-vulkan.cpp", "patch": "@@ -825,6 +825,8 @@ struct vk_device_struct {\n vk_pipeline pipeline_pool2d_f32;\n vk_pipeline pipeline_rwkv_wkv6_f32;\n vk_pipeline pipeline_rwkv_wkv7_f32;\n+ // [size_idx][kda] where size_idx: 0=d32, 1=d64, 2=d128\n+ vk_pipeline...
2026-03-12T10:32:04
golang/go
fe70866d1dc8c44ab19180ecab2b5c5b8628265a
e01a1c01f830e2398b773b803dce3238b1107ce9
runtime: throw on zero-sized range passed to addrRanges.add addrRanges represents a set of addresses. Currently, passing in a zero-sized range will cause that range to be added to the list, even though it doesn't represent any address (addrRanges.contains will still always return false, and findSucc will give surprisi...
[ { "path": "src/runtime/mranges.go", "patch": "@@ -218,7 +218,7 @@ func (a *addrRanges) contains(addr uintptr) bool {\n \n // add inserts a new address range to a.\n //\n-// r must not overlap with any address range in a.\n+// r must not overlap with any address range in a and r.size() must be > 0.\n func (a...
2020-07-15T18:56:39
huggingface/transformers
280cef51b37de3f04637ab94d7f34da8257131bc
2505357e4fdea82d1586b8c0862b4eeb46fac457
Give more useful `metric_for_best_model` errors (#31450) Give more useful metric_for_best_model errors
[ { "path": "src/transformers/trainer.py", "patch": "@@ -2749,7 +2749,13 @@ def _evaluate(self, trial, ignore_keys_for_eval, skip_scheduler=False):\n metric_to_check = self.args.metric_for_best_model\n if not metric_to_check.startswith(\"eval_\"):\n metric_to_check = f\...
2024-06-18T15:56:30
ggml-org/llama.cpp
c3e3f9e5338e8eb3d5b13d20f20d38b53d9d19ce
40c550d4f6838180fa40eda8351043353911ca40
convert : better mtp check and fix return [no ci] (#20419)
[ { "path": "convert_hf_to_gguf.py", "patch": "@@ -10092,9 +10092,9 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter\n # Skip Multi-Token Prediction (MTP) tensors. These are used for\n # for speculative decoding but we don't include them in this model...
2026-03-12T09:04:20
nodejs/node
61793dffe38411a2bbff5229fb9a4b3af771762d
f44f18a857a802367de1e38b7017bae7ab0dc44b
tools: speed up lint-md-build by using package-lock.json PR-URL: https://github.com/nodejs/node/pull/16945 Fixes: https://github.com/nodejs/node/issues/16628 Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com> Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com> Reviewed...
[ { "path": ".gitignore", "patch": "@@ -104,8 +104,8 @@ deps/npm/node_modules/.bin/\n \n # test artifacts\n tools/faketime\n-tools/remark-cli\n-tools/remark-preset-lint-node\n+tools/remark-cli/node_modules\n+tools/remark-preset-lint-node/node_modules\n icu_config.gypi\n *.tap\n ", "additions": 2, "del...
2017-11-10T22:31:46
huggingface/transformers
2505357e4fdea82d1586b8c0862b4eeb46fac457
4691ffbd41666057670de69f9e5b6fc445fb0fd9
Fix documentation typos (#31476) Fix doc typo
[ { "path": "src/transformers/models/cohere/configuration_cohere.py", "patch": "@@ -53,7 +53,7 @@ class CohereConfig(PretrainedConfig):\n num_key_value_heads (`int`, *optional*):\n This is the number of key_value heads that should be used to implement Grouped Query Attention. If\n ...
2024-06-18T15:09:50
ggml-org/llama.cpp
40c550d4f6838180fa40eda8351043353911ca40
de190154c85d20e24dbeae8c8af1849402ae5098
vulkan: fix SSM_CONV PP scaling with large ubatch sizes (#20379) * vulkan: optimize SSM_CONV workgroup dispatch for large ubatch Tile tokens into 2D workgroups (32x16) to reduce workgroup launch overhead at large ubatch sizes. Add vec4 fast path for nc=4 (common d_conv size). Fixes PP performance degradation with uba...
[ { "path": "ggml/src/ggml-vulkan/ggml-vulkan.cpp", "patch": "@@ -4576,7 +4576,7 @@ static void ggml_vk_load_shaders(vk_device& device) {\n ggml_vk_create_pipeline(device, device->pipeline_ssm_scan_f32_d256, \"ssm_scan_256_f32\", ssm_scan_f32_len, ssm_scan_f32_data, \"main\", 8, sizeof(vk_op_ssm_scan_...
2026-03-12T09:03:18
golang/go
9db7db54b0e84d6b3ace94cb1f2a42e065575f17
a972e8ccb9824d97f029bde5409527fb27e85989
log: expose std via new Default function To allow passing around the package level *Logger, it is now exposed to callers of the Default function. We considered exposing std, however at this time there is no need to allow callers to replace std only pass and call methods directly. Fixes #39057 Change-Id: I710b16a3aa5...
[ { "path": "src/log/log.go", "patch": "@@ -75,6 +75,9 @@ func (l *Logger) SetOutput(w io.Writer) {\n \n var std = New(os.Stderr, \"\", LstdFlags)\n \n+// Default returns the *Logger used by the package-level output functions.\n+func Default() *Logger { return std }\n+\n // Cheap integer to fixed-width decima...
2020-10-22T22:16:01
nodejs/node
803cacd228c273214072a6b9ca98d8c83d217c97
e7f30db1a682208f2e5026ee299f07a3f23958e3
test: use fixtures module for path resolve PR-URL: https://github.com/nodejs/node/pull/16842 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com> Reviewed-By: Luigi Pinca <luigipinca@gmail.com> Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Colin Ihrig <cjihrig@gma...
[ { "path": "test/parallel/test-http2-respond-with-fd-errors.js", "patch": "@@ -1,10 +1,13 @@\n 'use strict';\n \n const common = require('../common');\n+\n if (!common.hasCrypto)\n common.skip('missing crypto');\n+\n+const fixtures = require('../common/fixtures');\n+\n const http2 = require('http2');\n-con...
2017-11-06T16:45:07
ollama/ollama
c2d813bdc3665ffad7618b726e370da5f0f3795a
786f3a1c44fb3733f6f1b5e7a5d20d6d772a4e47
Fix rocm deps with new subprocess paths
[ { "path": "Dockerfile", "patch": "@@ -42,7 +42,7 @@ ARG CGO_CFLAGS\n ARG AMDGPU_TARGETS\n RUN OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh\n RUN mkdir /tmp/scratch && \\\n- for dep in $(cat /go/src/github.com/ollama/ollama/llm/llama.cpp/build/linux/x86_64/rocm*/lib/deps.txt) ; do \\\n+ for dep in $(zcat...
2024-04-11T19:52:06
huggingface/transformers
6e56b83453c8f376f6a1d4442ae51896929b6df7
28316d0e8b4726c39bf75572c7788792c7da9ce9
Update chat template docs and bump Jinja version (#31455) * Update chat template docs * Minor bug in the version check * Update docs/source/en/chat_templating.md Co-authored-by: Joshua Lochner <admin@xenova.com> * Update docs/source/en/chat_templating.md Co-authored-by: Joshua Lochner <admin@xenova.com>...
[ { "path": "docs/source/en/chat_templating.md", "patch": "@@ -573,23 +573,21 @@ default template for that model class is used instead. Let's take a look at the\n \"{% for message in messages %}{% if message['role'] == 'user' %}{{ ' ' }}{% endif %}{{ message['content'] }}{% if not loop.last %}{{ ' ' }}{% end...
2024-06-18T13:16:30
ggml-org/llama.cpp
05039967da783fa8c7b8bc51e80ae834ab1cda73
e4cff0956b0dd7196cd8d4717daf02329cdeb870
ggml-virtgpu: Fix some build commands (#20341)
[ { "path": "docs/backend/VirtGPU/development.md", "patch": "@@ -55,7 +55,8 @@ LLAMA_MAC_BUILD=$PWD/build/ggml-virtgpu-backend\n cmake -S . -B $LLAMA_MAC_BUILD \\\n -DGGML_NATIVE=OFF \\\n -DLLAMA_CURL=ON \\\n- -DGGML_REMOTINGBACKEND=ONLY \\\n+ -DGGML_VIRTGPU=ON \\\n+ -DGGML_VIRTGPU_...
2026-03-12T07:47:45
golang/go
a972e8ccb9824d97f029bde5409527fb27e85989
db185e543fe471c522790b7d93291e786dc54a84
all: fix quoting for compress/bzip2 and time's godoc The existing usage of grave accent (`) and apostrophe (') at some places made godoc to ignore them and show it as it is. So, use both of the characters twice (consecutively) so that godoc can convert it to {left,right} double quotation mark. Fixes #41958 Change-Id...
[ { "path": "src/compress/bzip2/bzip2.go", "patch": "@@ -29,8 +29,8 @@ type reader struct {\n \tsetupDone bool // true if we have parsed the bzip2 header.\n \tblockSize int // blockSize in bytes, i.e. 900 * 1000.\n \teof bool\n-\tc [256]uint // the `C' array for the inverse BWT.\n-\...
2020-10-14T18:54:14
nodejs/node
e7f30db1a682208f2e5026ee299f07a3f23958e3
2b9315199d319cd5d1afc1a5bf959b610bd343a9
deps: cherry-pick 3c8195d from V8 upstream Original commit message: [map] Fix map constructor to correctly throw. We need to throw before rethrowing, otherwise the exception does not trigger a debugger event and is not reported if uncaught. R=gsathya@chromium.org, jgruber@chromium.org Bug: v8:7047 Chan...
[ { "path": "common.gypi", "patch": "@@ -27,7 +27,7 @@\n \n # Reset this number to 0 on major V8 upgrades.\n # Increment by one for each non-official patch applied to deps/v8.\n- 'v8_embedder_string': '-node.10',\n+ 'v8_embedder_string': '-node.11',\n \n # Enable disassembler for `--print-co...
2017-11-09T04:58:41
ollama/ollama
a7b431e7435aeacd084c0e3d59d8d035b6a667b6
5a25f935223e458e1c1042e00613379cd1364674
server: provide helpful workaround hint when stalling on pull (#3584) This is a quick fix to help users who are stuck on the "pull" step at 99%. In the near future we're introducing a new registry client that should/will hopefully be smarter. In the meantime, this should unblock the users hitting issue #1736.
[ { "path": "server/download.go", "patch": "@@ -247,7 +247,8 @@ func (b *blobDownload) downloadChunk(ctx context.Context, requestURL *url.URL, w\n \t\t\t\t}\n \n \t\t\t\tif !part.lastUpdated.IsZero() && time.Since(part.lastUpdated) > 5*time.Second {\n-\t\t\t\t\tslog.Info(fmt.Sprintf(\"%s part %d stalled; retr...
2024-04-10T23:24:37
ggml-org/llama.cpp
e4cff0956b0dd7196cd8d4717daf02329cdeb870
4cc6eb158c83a3b9833526f3bf1175696884fb0c
metal : avoid divisions in bin kernel (#20426) * metal : avoid modulus in bin kernel when not broadcasting * metal : fix capture_started flag
[ { "path": "ggml/src/ggml-metal/ggml-metal-context.m", "patch": "@@ -554,7 +554,7 @@ enum ggml_status ggml_metal_graph_compute(ggml_metal_t ctx, struct ggml_cgraph *\n \n // enter here only when capturing in order to wait for all computation to finish\n // otherwise, we leave the graph to com...
2026-03-12T07:42:40
huggingface/transformers
28316d0e8b4726c39bf75572c7788792c7da9ce9
dabf01973acdc9b0f28a5366c5693f3ffff9942d
Fix single letter stop strings (#31448) * Fix single letter stop strings * Change the 0 to a 1 to avoid potential empty vector headaches later * Restructure for clarity * Update tests/generation/test_stopping_criteria.py Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com> * Add t...
[ { "path": "src/transformers/generation/stopping_criteria.py", "patch": "@@ -372,10 +372,11 @@ def _stop_string_create_embedding_vec(token_list, token_indices, stop_strings) -\n token_valid_positions, token_end_overlaps = StopStringCriteria._stop_string_get_matching_positions(\n token_lis...
2024-06-18T13:07:16
golang/go
f24ff3856a629a6b5fefe28a1676638d5f103342
4fdb98dcb1bf7a33565c81db3c00e91a466e098d
cmd/go: change error message for missing import with unused replacement In readonly mode, if a package is not provided by any module in the build list, and there is an unused replacement that contains the package, we now recommend a 'go get' command to add a requirement on the highest replaced version. Fixes #41416 ...
[ { "path": "src/cmd/go/internal/modfetch/pseudo.go", "patch": "@@ -76,6 +76,12 @@ func PseudoVersion(major, older string, t time.Time, rev string) string {\n \treturn v + incDecimal(patch) + \"-0.\" + segment + build\n }\n \n+// ZeroPseudoVersion returns a pseudo-version with a zero timestamp and\n+// revisi...
2020-10-16T18:02:16
nodejs/node
2b9315199d319cd5d1afc1a5bf959b610bd343a9
c07d1279b230fc2c178c748b998979cca374c86b
src: fix compiler warning in process.ppid This commit fixes a compiler warning with SetAccessor() by passing in the context. PR-URL: https://github.com/nodejs/node/pull/16958 Reviewed-By: Anna Henningsen <anna@addaleax.net> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com> Reviewed-By: Timothy Gu <timothygu99@gmail...
[ { "path": "src/node.cc", "patch": "@@ -3267,8 +3267,9 @@ void SetupProcessObject(Environment* env,\n READONLY_PROPERTY(process, \"pid\", Integer::New(env->isolate(), getpid()));\n READONLY_PROPERTY(process, \"features\", GetFeatures(env));\n \n- process->SetAccessor(FIXED_ONE_BYTE_STRING(env->isolate()...
2017-11-12T02:10:13
ggml-org/llama.cpp
246ffc4b057504712272433b035afefc78457cda
aa429cf5075ef1cf14c9cf784c73fc436ff6eede
vulkan: fix l2_norm epsilon handling (#20350)
[ { "path": "ggml/src/ggml-vulkan/ggml-vulkan.cpp", "patch": "@@ -16061,7 +16061,7 @@ static void ggml_vk_check_results_0(ggml_backend_vk_context * ctx, ggml_cgraph *\n tensor_clone = ggml_arange(ggml_ctx, start, stop, step);\n } else if (tensor->op == GGML_OP_FILL) {\n const f...
2026-03-12T05:39:41
huggingface/transformers
cd71f9381b86b0dc1fd60e8b87fb5bade35aa0cd
76289fbc7cb4d205483dce38bd7b12975d9f3195
Donut: fix `generate` call from local path (#31470) * local donut path fix * engrish * Update src/transformers/generation/utils.py Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com> --------- Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
[ { "path": "src/transformers/generation/utils.py", "patch": "@@ -575,8 +575,12 @@ def _prepare_decoder_input_ids_for_generation(\n # no user input -> use decoder_start_token_id as decoder_input_ids\n if decoder_input_ids is None:\n decoder_input_ids = decoder_start_token_id\n- ...
2024-06-18T12:28:06
golang/go
4fdb98dcb1bf7a33565c81db3c00e91a466e098d
5cd4390f3853b8d0d2d962f7acdac87c0eba3d77
cmd/go: save sums for zips needed to diagnose ambiguous imports Previously, we would retain entries in go.sum for .mod files in the module graph (reachable from the main module) and for .zip files of modules providing packages. This isn't quite enough: when we load a package, we need the content of each module in the...
[ { "path": "src/cmd/go/internal/modload/init.go", "patch": "@@ -18,6 +18,7 @@ import (\n \t\"path/filepath\"\n \t\"strconv\"\n \t\"strings\"\n+\t\"sync\"\n \n \t\"cmd/go/internal/base\"\n \t\"cmd/go/internal/cfg\"\n@@ -911,7 +912,10 @@ func WriteGoMod() {\n \t\t// The go.mod file has the same semantic conten...
2020-10-15T22:22:09
ggml-org/llama.cpp
aa429cf5075ef1cf14c9cf784c73fc436ff6eede
5866e3bbc8a2ede4236871ea73a3eb3337abf530
vulkan: fix OOB check in flash_attn_mask_opt (#20296)
[ { "path": "ggml/src/ggml-vulkan/ggml-vulkan.cpp", "patch": "@@ -8840,7 +8840,7 @@ static void ggml_vk_flash_attn(ggml_backend_vk_context * ctx, vk_context& subctx\n }\n \n // Only use mask opt when the mask is fairly large. This hasn't been tuned extensively.\n- bool use_mask_opt = mask && nem1 >...
2026-03-12T05:35:49
ollama/ollama
2b4ca6cf36c0dc31fdae7046433f4341f171026f
ad90b9ab3d94e330970b00ae29be50b6ed62a8fb
fix ci
[ { "path": ".github/workflows/test.yaml", "patch": "@@ -5,7 +5,6 @@ on:\n paths:\n - '**/*'\n - '!docs/**'\n- - '!examples/**'\n - '!README.md'\n \n jobs:", "additions": 0, "deletions": 1, "language": "YAML" }, { "path": "app/tray/tray.go", "patch": "@@ -24,...
2024-04-10T18:26:15
nodejs/node
29de91b64f2ec18e9d91fb89f58f371452f7b76e
0e10717e437eeb05e57d86e823777f418f00c478
doc: fix typo in http2 doc `Sesssion` -> `Session` PR-URL: https://github.com/nodejs/node/pull/16993 Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com> Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Anatoli Papirovski <apapirovski@mac.com> Reviewed-By: Luigi Pinca <luigipinca@gmail.com> Reviewed-By: James...
[ { "path": "doc/api/http2.md", "patch": "@@ -821,7 +821,7 @@ Shortcut for `http2stream.rstStream()` using error code `0x02` (Internal Error).\n added: v8.4.0\n -->\n \n-* Value: {Http2Sesssion}\n+* Value: {Http2Session}\n \n A reference to the `Http2Session` instance that owns this `Http2Stream`. The\n value...
2017-11-13T15:16:29
huggingface/transformers
b38612d31263294bc997a21e4aceb1485414fa2d
1f9387d33d40562add4be03176e22ddf9ed0d4ab
Agents: Improve python interpreter (#31409) * Improve Python interpreter * Add with and assert statements * Prevent overwriting existing tools * Check interpreter errors are well logged in code agent * Add lazy evaluation for and and or * Improve variable assignment * Fix early return statements in functions * ...
[ { "path": "src/transformers/agents/default_tools.py", "patch": "@@ -34,11 +34,16 @@ def custom_print(*args):\n \n BASE_PYTHON_TOOLS = {\n \"print\": custom_print,\n+ \"isinstance\": isinstance,\n \"range\": range,\n \"float\": float,\n \"int\": int,\n \"bool\": bool,\n \"str\": st...
2024-06-18T09:55:36
golang/go
5cd4390f3853b8d0d2d962f7acdac87c0eba3d77
5f616a6fe789622f3e0ed0e8a00db9471e2a02f4
cmd/go: don't fetch files missing sums in readonly mode If the go command needs a .mod or .zip file in -mod=readonly mode (now the default), and that file doesn't have a hash in the main module's go.sum file, the go command will now report an error before fetching the file, rather than at the end when failing to updat...
[ { "path": "src/cmd/go/internal/load/pkg.go", "patch": "@@ -254,8 +254,8 @@ func (p *Package) setLoadPackageDataError(err error, path string, stk *ImportSta\n \t// package's source files themselves (scanner errors).\n \t//\n \t// TODO(matloob): Perhaps make each of those the errors in the first group\n-\t// ...
2020-10-13T22:19:21
ggml-org/llama.cpp
5866e3bbc8a2ede4236871ea73a3eb3337abf530
0516e04bf9c002bf190519af6795415ab3ebd059
vulkan: Fix ErrorOutOfHostMemory on Intel GPU when loading large models with --no-mmap (#20059) * Changed to reuse command buffers to fix crashing on Intel GPU * Removed unused parameter * Fixed compile error and minor mistake * Fix logging * Changing to use usage flag per command buffer * fixed style * added bu...
[ { "path": "ggml/src/ggml-vulkan/ggml-vulkan.cpp", "patch": "@@ -27,6 +27,7 @@ DispatchLoaderDynamic & ggml_vk_default_dispatcher();\n #include <iostream>\n #include <tuple>\n #include <vector>\n+#include <deque>\n #include <sstream>\n #include <utility>\n #include <memory>\n@@ -188,17 +189,28 @@ struct ggml...
2026-03-12T05:30:16
ollama/ollama
ad90b9ab3d94e330970b00ae29be50b6ed62a8fb
4340f8eba4b3a0e5672a0c4837ac45a23c074fdf
api: start adding documentation to package api (#2878) * api: start adding documentation to package api Updates #2840 * Fix lint typo report
[ { "path": "api/client.go", "patch": "@@ -1,3 +1,9 @@\n+// Package api implements the client-side API for code wishing to interact\n+// with the ollama service. The methods of the [Client] type correspond to\n+// the ollama REST API as described in https://github.com/ollama/ollama/blob/main/docs/api.md\n+//\...
2024-04-10T17:31:55
nodejs/node
0e10717e437eeb05e57d86e823777f418f00c478
14d24cc4d226a8317ea4eb1938840f86588dddfe
build: runtime-deprecate requiring deps PR-URL: https://github.com/nodejs/node/pull/16392 Fixes: https://github.com/nodejs/node/pull/15566#discussion_r146111170 Reviewed-By: Refael Ackermann <refack@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Ben ...
[ { "path": "doc/api/deprecations.md", "patch": "@@ -747,6 +747,37 @@ be set to `false` to disable ECDH entirely on the server only. This mode is\n deprecated in preparation for migrating to OpenSSL 1.1.0 and consistency with\n the client. Use the `ciphers` parameter instead.\n \n+<a id=\"DEP0084\"></a>\n+###...
2017-10-22T19:16:48
huggingface/transformers
1f9387d33d40562add4be03176e22ddf9ed0d4ab
9ba9369a2557e53a01378199a9839ec6e82d8bc7
Fix typing errors in `Qwen2ForTokenClassification` (#31440) * Update modeling_qwen2.py * Fix llama * More fixes
[ { "path": "src/transformers/models/llama/modeling_llama.py", "patch": "@@ -1545,7 +1545,7 @@ def set_input_embeddings(self, value):\n @add_start_docstrings_to_model_forward(LLAMA_INPUTS_DOCSTRING)\n def forward(\n self,\n- input_ids: torch.LongTensor = None,\n+ input_ids: Optio...
2024-06-18T09:27:18
golang/go
5f616a6fe789622f3e0ed0e8a00db9471e2a02f4
d05e89a8fd35bb543df6a29faea81a85565db92f
cmd/go: in 'go mod init', suggest running 'go mod tidy' When 'go mod init' is run in an existing project, it may import requirements from a vendor configuration file, but the requirements may not be complete, and go.sum won't contain sums for module zips. With -mod=readonly, the next build command is likely to fail. ...
[ { "path": "src/cmd/go/internal/modload/init.go", "patch": "@@ -435,6 +435,29 @@ func CreateModFile(ctx context.Context, modPath string) {\n \n \tmodFileToBuildList()\n \tWriteGoMod()\n+\n+\t// Suggest running 'go mod tidy' unless the project is empty. Even if we\n+\t// imported all the correct requirements ...
2020-10-23T15:10:10
ggml-org/llama.cpp
3d9ab225e71d96458adc2219f0d039bc46dbc8b9
d63aa398de455cf8243802c5099dc30a2e26edcc
opencl: add cumsum op (#18981) * OpenCL: add CUMSUM op support * remove unused argument * opencl: refactor cumsum * opencl: refactor * opencl: refactor tmp buffer * opencl: adjust max number of subgroups * opencl: fix whitespace * opencl: fix global size when cumsum the tmp buffer --------- Co-authored-by: Li...
[ { "path": "ggml/src/ggml-opencl/CMakeLists.txt", "patch": "@@ -132,6 +132,7 @@ set(GGML_OPENCL_KERNELS\n ssm_conv\n sub\n sum_rows\n+ cumsum\n transpose\n concat\n tsembd", "additions": 1, "deletions": 0, "language": "Plain Text" }, { "path": "ggml/src/ggml-ope...
2026-03-12T05:03:07
rust-lang/rust
f68fd669f3d7a0847b5c2f3b789b6c95db0acdc1
c4f727bfb8640e9ddd4f66b562bba7a0d7b4f46c
Fix logical error in relevance scoring implementation
[ { "path": "src/tools/rust-analyzer/crates/ide-completion/src/item.rs", "patch": "@@ -252,14 +252,14 @@ impl CompletionRelevance {\n /// Provides a relevance score. Higher values are more relevant.\n ///\n /// The absolute value of the relevance score is not meaningful, for\n- /// example a va...
2025-03-07T05:24:14
nodejs/node
1601a3c2a9d82c0c235bc37c5ecbb365a4ceb47c
107e8a9867bc47e1cb7d16bfc9f5efcf300dee82
test: used fixturesDir from fixtures modules In test-fs-realpath-on-substed-drive, require common/fixtures module and swapped the location of fixturesDir from common to fixtures module. PR-URL: https://github.com/nodejs/node/pull/16813 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Rich Trott <rtrott@gma...
[ { "path": "test/parallel/test-fs-realpath-on-substed-drive.js", "patch": "@@ -4,6 +4,8 @@ const common = require('../common');\n if (!common.isWindows)\n common.skip('Test for Windows only');\n \n+const fixtures = require('../common/fixtures');\n+\n const assert = require('assert');\n const fs = require('...
2017-11-06T15:05:59
huggingface/transformers
9ba9369a2557e53a01378199a9839ec6e82d8bc7
02300273e220932a449a47ebbe453e7789be454b
simple fix (#31456)
[ { "path": "examples/pytorch/text-classification/run_classification.py", "patch": "@@ -87,7 +87,7 @@ class DataTrainingArguments:\n },\n )\n text_column_delimiter: Optional[str] = field(\n- default=\" \", metadata={\"help\": \"THe delimiter to use to join text columns into a single sen...
2024-06-17T21:30:37
ggml-org/llama.cpp
d63aa398de455cf8243802c5099dc30a2e26edcc
a8304b4d279cadc240802388d1532bea650de4c8
hip: compile debug builds with -O2 on hip to avoid a compiler bug (#20392)
[ { "path": "ggml/src/ggml-hip/CMakeLists.txt", "patch": "@@ -11,6 +11,10 @@ endif()\n list(APPEND CMAKE_PREFIX_PATH ${ROCM_PATH})\n list(APPEND CMAKE_PREFIX_PATH \"${ROCM_PATH}/lib64/cmake\")\n \n+if (NOT DEFINED CMAKE_HIP_FLAGS_DEBUG)\n+ set(CMAKE_HIP_FLAGS_DEBUG \"-g -O2\")\n+endif()\n+\n # CMake on Wi...
2026-03-12T02:37:10
golang/go
d05e89a8fd35bb543df6a29faea81a85565db92f
c8c3c29daa74f2d3e1a26f2e289ad3d2b9ba20dd
cmd/go: refactor modload.InitMod InitMod is split into two functions. LoadModFile parses an existing go.mod file and loads the build list (or checks vendor/modules.txt for consistency in vendor mode). CreateModFile creates a new go.mod file, possibly inferring the module path and importing a vendor configuration file....
[ { "path": "src/cmd/go/internal/list/list.go", "patch": "@@ -415,7 +415,7 @@ func runList(ctx context.Context, cmd *base.Command, args []string) {\n \t\t\tbase.Fatalf(\"go list -m: not using modules\")\n \t\t}\n \n-\t\tmodload.InitMod(ctx) // Parses go.mod and sets cfg.BuildMod.\n+\t\tmodload.LoadModFile(ctx...
2020-10-22T22:20:00
ollama/ollama
01114b45265dff484fd56d0430317c32dd4d9bd4
1524f323a3392adc2613b066b2b1c1e46887001d
fix: rope
[ { "path": "api/types.go", "patch": "@@ -122,6 +122,11 @@ type Runner struct {\n \tUseMMap bool `json:\"use_mmap,omitempty\"`\n \tUseMLock bool `json:\"use_mlock,omitempty\"`\n \tNumThread int `json:\"num_thread,omitempty\"`\n+\n+\t// Unused: RopeFrequencyBase is ignored. Instead the value in the model w...
2024-04-09T23:15:24
nodejs/node
f5809d8e35dbbd826ce72a225cd5694dc9252e31
1aac4c1d8f0860d8050baabaef255508d001114b
build: fix cctest compilation Currently the cctest target compiles sources files even though they are compiled for the node target. This is my fault as when I worked on the task of getting the cctest to use the object files from the node target I missed a few sources that were being included from node.gypi. This also ...
[ { "path": "doc/guides/writing-tests.md", "patch": "@@ -325,6 +325,11 @@ Next add the test to the `sources` in the `cctest` target in node.gyp:\n ...\n ],\n ```\n+Note that the only sources that should be included in the cctest target are\n+actual test or helper source files. There might be a need to inclu...
2017-11-08T08:41:09
ggml-org/llama.cpp
fdb17643d379cd35bf6acf0f57cfaa500f88a145
1eea6a2968b401c8377bc9d4d49a7a64ffbfb3d0
model : add support for Phi4ForCausalLMV (#20168) * Add support for Phi4ForCausalLMV. * Fix Phi-4 vision parity (correcting SigLIP2 patch-kernel export layout) and matching HF NaFlex resize behavior in mtmd. * Rename contants + fix tokenizer label * Clean-ups. * Fix GGUF export. * Set tokenizer.ggml.pre explicitl...
[ { "path": "convert_hf_to_gguf.py", "patch": "@@ -5062,7 +5062,7 @@ def set_gguf_parameters(self):\n self.gguf_writer.add_add_bos_token(False)\n \n \n-@ModelBase.register(\"Phi3ForCausalLM\")\n+@ModelBase.register(\"Phi3ForCausalLM\", \"Phi4ForCausalLMV\")\n class Phi3MiniModel(TextModel):\n mode...
2026-03-11T23:25:54