id int64 393k 2.82B | repo stringclasses 68
values | title stringlengths 1 936 | body stringlengths 0 256k ⌀ | labels stringlengths 2 508 | priority stringclasses 3
values | severity stringclasses 3
values |
|---|---|---|---|---|---|---|
2,347,931,743 | langchain | ChatGoogleGenerativeAI does not support streaming | ### Checked other resources
- [X] I added a very descriptive title to this issue.
- [X] I searched the LangChain documentation with the integrated search.
- [X] I used the GitHub search to find a similar question and didn't find it.
- [X] I am sure that this is a bug in LangChain rather than my code.
- [X] The bug is ... | 🤖:bug,🔌: google | low | Critical |
2,347,938,581 | pytorch | DISABLED test_mixed_mm_bad_cases (__main__.TestPatternMatcher) | Platforms: rocm
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_mixed_mm_bad_cases&suite=TestPatternMatcher&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/26107149296).
Over the past 3 hours, it has bee... | module: rocm,triaged,module: flaky-tests,skipped,module: inductor | low | Critical |
2,347,952,731 | pytorch | nvcc -broken when use Libtorch | I want to use libtorch in windows10, I met the following problem during cmake.
build# cmake ..
-- Building for: MinGW Makefiles
-- The C compiler identification is GNU 8.1.0
-- The CXX compiler identification is GNU 8.1.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for w... | needs reproduction,module: windows,module: cpp,triaged | low | Critical |
2,347,973,132 | ollama | support for recurrent gemma | https://huggingface.co/google/recurrentgemma-2b-it
Support for recurrent gemma | model request | low | Major |
2,348,008,604 | ollama | Apple Silicon macs with 8GB or 16GB slow down when loading larger models | ### What is the issue?
Less of the model should be loaded to Metal to avoid causing lag
### OS
macOS
### GPU
_No response_
### CPU
_No response_
### Ollama version
_No response_ | bug | low | Major |
2,348,034,332 | pytorch | Improve type inference in torchscript | ### 🚀 The feature, motivation and pitch
Tracing a simple function fails due to lacking type information
```python
import torch
def f(x, i=-1):
return x[i]
torch.jit.script(f)
```
currently gives the following error
```console
RuntimeError:
Expected a default value of type Tensor (inferred) on... | oncall: jit | low | Critical |
2,348,088,232 | TypeScript | Nested generic calls don't inherit outer inferences when their return types are intersections containing functions | ### 🔎 Search Terms
generic call inference intersection return type
### 🕗 Version & Regression Information
- This is the behavior in every version I tried
### ⏯ Playground Link
https://www.typescriptlang.org/play/?ts=5.6.0-dev.20240611#code/JYOwLgpgTgZghgYwgAgMIHsQ2AcwDwAqAfMgN4CwAUMsgppAB5gBcyBA3FTaAG7oDWEVhWo1k... | Help Wanted,Possible Improvement | low | Critical |
2,348,119,051 | pytorch | Inconsistency `CrossEntropyLoss` vs `BCELoss` regarding logits/probability space | ### 🚀 The feature, motivation and pitch
The loss [`CrossEntropyLoss`](https://pytorch.org/docs/stable/generated/torch.nn.CrossEntropyLoss.html#torch.nn.CrossEntropyLoss) expects logits inputs. On the other hand, there exist [`BCELoss`](https://pytorch.org/docs/stable/generated/torch.nn.BCELoss.html#torch.nn.BCELoss... | module: nn,triaged | low | Minor |
2,348,124,666 | node | NodeJS unconditionally uses AVX512 instructions even if they are disabled | ### Version
v22.0.0-pre and above
### Platform
Linux ZenTower 6.8.0-35-generic #35-Ubuntu SMP PREEMPT_DYNAMIC Mon May 20 15:51:52 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
### Subsystem
base64
### What steps will reproduce the bug?
On a linux machine which supports the AVX-512 extension, disable it, then run `./ou... | dependencies | low | Critical |
2,348,143,577 | pytorch | 'make html' will print 'duplicate object description' warnings when there are 1~5 CPUs in the running machine | ### 🐛 Describe the bug
Hello, PyTorch team.
I discovered that when building the PyTorch documentation using `sphinx-build`, if the `sphinx-build` command does not include the `-j auto` argument, the following warning message appears:
```bash
/home/hwhsu1231/Repo/testing/pytorch-v2.3.1/docs/source/nn.rst:566:... | module: docs,triaged | low | Critical |
2,348,146,588 | pytorch | torch.jit.script not work for ParameterDict.items | ### 🐛 Describe the bug
As it says in the title, using `ParameterDict.items()` will break `torch.jit.script` as seen below
```
import torch
from torch import nn
class MyModule(nn.Module):
def __init__(self):
super().__init__()
self.parameter_dict = nn.ParameterDict(
{str(i): ... | oncall: jit | low | Critical |
2,348,151,931 | godot | GPUparticle rigidbody collision snap delay | ### Tested versions
-Reproducible in 4.2.1 and 4.3 and possibly earlier versions
### System information
Godot v4.2.1.stable - Windows 10.0.22631 - Vulkan (Forward+) - dedicated NVIDIA GeForce RTX 3050 Laptop GPU (NVIDIA; 31.0.15.4633) - 11th Gen Intel(R) Core(TM) i5-11300H @ 3.10GHz (8 Threads)
### Issue de... | bug,topic:particles | low | Major |
2,348,179,667 | vscode | Lecagy Linux server used in WSL under Ubuntu 22.04 | Version: 1.91.0-insider (user setup)
Commit: 51d598183ea632c658c34e0a11da5706192a1b69
Date: 2024-06-12T05:49:43.246Z
Electron: 29.4.0
ElectronBuildId: 9593362
Chromium: 122.0.6261.156
Node.js: 20.9.0
V8: 12.2.281.27-electron.0
OS: Windows_NT x64 10.0.22635
WSL
```
lsb_release -a
No LSB modules are available.
Distribut... | bug,remote,confirmation-pending,electron-29-update | low | Minor |
2,348,233,803 | transformers | Problems when using SinkCache for model.generate() | ### System Info
- `transformers` version: 4.41.2
- Platform: Linux-5.4.0-173-generic-x86_64-with-glibc2.31
- Python version: 3.10.0
- Huggingface_hub version: 0.23.3
- Safetensors version: 0.4.3
- Accelerate version: 0.31.0
- Accelerate config: not found
- PyTorch version (GPU?): 2.3.1+cu121 (True)
- Tensor... | WIP,Generation,Cache | low | Critical |
2,348,236,868 | pytorch | [bug]When the variable in the dataset is of type cuda, calling the dataloader will result in 0 | ### 🐛 Describe the bug
```python
import torch
from torch.utils.data import Dataset
# from datasets.base_dataset import BaseDataset
class dataset(Dataset):
def __init__(self):
self.class_weights = torch.tensor([0.8373, 0.918, 0.866, 1.0345,
1.0166, 0.9969... | module: dataloader,triaged,module: data | low | Critical |
2,348,247,986 | neovim | send alt-x keycodes to vim.fn.chansend() | ### Problem
I've been tinkering around with nested neovim instances and sending pre-recorded keystrokes to them with `vim.fn.feedkeys`. In order to allow me to send these keys without having to have the nested sessions open in a terminal as my current window, I've been trying to transition my code to use `vim.fn.cha... | enhancement,terminal | low | Major |
2,348,289,448 | flutter | Flutter's semantics search has non-linear code in it that can be super slow | When running a large app (DevTools on linux) with `flutter run -d linux` I noticed that the app freezes. When l profiled it via DevTools I noticed it spends a lot of time in the semantics code.
`_drawFrame` -> `_handleDrawFrame` -> ... `RendererBinding.drawFrame` -> ... -> `PipelineOwner.flushSemantics` -> ... -> `S... | framework,a: accessibility,P2,team-framework,triaged-framework | low | Major |
2,348,301,599 | godot | Unexpected scene tile recreation after calling `update_internals()` | ### Tested versions
4.2.2.stable / 4.3.beta1
### System information
Arch Linux
### Issue description
It's documented that scene tiles may not be available when the tilemap is ready, and users can call `update_internals()` to force immediate update.
https://github.com/godotengine/godot/blob/3a11eb98181d8fa0b173a... | bug,topic:2d | low | Critical |
2,348,322,861 | opencv | opencv still uses Concurrency Runtime in parallel_for_xxx functions on Windows | ### Describe the feature and motivation
Concurrency Runtime is still used by opencv in the implementation of [parallel_for_xxx](https://github.com/opencv/opencv/blob/4.x/modules/core/src/parallel.cpp) functions on Windows. We have recently filed a [bug report on Concurrency Runtime task scheduling](https://developerco... | feature,platform: win32 | low | Critical |
2,348,323,584 | PowerToys | Remap a shortcut is not working when targeting git-bash | ### Microsoft PowerToys version
0.81.1
### Installation method
PowerToys auto-update
### Running as admin
Yes
### Area(s) with issue?
Keyboard Manager
### Steps to reproduce
1. Create a shortcut From `Shift+a` to `b` targeting `git-bash.exe`
2. Open git-bash and press `Shift+a`
### ✔️ Expected Behavior
`b` ... | Issue-Bug,Needs-Triage | low | Minor |
2,348,357,933 | godot | USER ERROR: FATAL: Index p_index = 4294967294 is out of bounds (count = 32772). at: operator[] (./core/templates/local_vector.h:161) | ### Tested versions
- Reproducable, though I'm sure this is a rare instance lol.
Godot 4.2.2, Forward+
### System information
Godot v4.2.2.stable.mono - Windows 10.0.19045 - Vulkan (Forward+) - dedicated AMD Radeon RX 7900 XT (Advanced Micro Devices, Inc.; 31.0.24002.92) - 13th Gen Intel(R) Core(TM) i5-13600K (20... | bug,needs testing,crash | low | Critical |
2,348,399,374 | vscode | Active line number not bold when word wrap is enabled | <!-- ⚠️⚠️ Do Not Delete This! bug_report_template ⚠️⚠️ -->
<!-- Please read our Rules of Conduct: https://opensource.microsoft.com/codeofconduct/ -->
<!-- 🕮 Read our guide about submitting issues: https://github.com/microsoft/vscode/wiki/Submitting-Bugs-and-Suggestions -->
<!-- 🔎 Search existing issues to avoid cr... | bug,editor-wrapping | low | Critical |
2,348,462,590 | pytorch | torch.onnx.export with dynamic axes fails for torch.nn.InstanceNorm1d with track_running_stats=True | ### 🐛 Describe the bug
`torch.onnx.export` fails for `torch.nn.InstanceNorm1d` (and 2d for that sake) when initialized with `track_running_stats=True` and using dynamic batch size.
Minimal reproducing code:
```python
import torch
B, C, N = 3, 7, 11
model = torch.nn.InstanceNorm1d(C, track_running_stats=Tr... | module: onnx,triaged,OSS contribution wanted | low | Critical |
2,348,479,603 | angular | Mark pure pipes dirty when a signal that they read changes | ### Which @angular/* package(s) are relevant/related to the feature request?
core
### Description
Currently, if the `transform` method of a pure pipe reads a signal, the pipe won't get reexecuted if the signal changes.
See https://stackblitz.com/edit/stackblitz-starters-eadv1h?file=src%2Fmain.ts as an examp... | area: core,core: pipes,core: reactivity,cross-cutting: signals | medium | Major |
2,348,531,457 | pytorch | crash@sleef_tryVXE2 () while trying to run torch.compile() BERT model | ### 🐛 Describe the bug
Hi,
I have been trying to load a pre-trained bert model `bert-base-cased` downloaded from HuggingFace and optimize it with torch.compile() on linux-s390x. It crashed -
```python
0x000003ffb91c4418 in sleef_tryVXE2 () from /root/anaconda3/envs/pytorch_test/lib/python3.10/site-package... | module: crash,triaged,module: sleef,module: POWER | low | Critical |
2,348,562,921 | rust | CURRENT_RUSTC_VERSION does not work with `#[deprecated]` | Found by @m-ou-se [here](https://github.com/rust-lang/rust/pull/126322#issuecomment-2162761978):
```
Compiling libc v0.2.155
Compiling memchr v2.5.0
Compiling std v0.0.0 (/checkout/library/std)
Compiling compiler_builtins v0.1.109
error: 'since' must be a Rust version number, such as "1.31.0"
|
... | A-attributes,T-libs,requires-internal-features,L-deprecated | low | Critical |
2,348,605,461 | godot | C# scenes with `NodePath`s randomly fail to load correctly (instead replaced with null) | ### Tested versions
- Reproducible in 4.2.2.stable.mono
- Not reproducible in 3.5
### System information
Most likely Windows 10 with Vulkan Forward+ renderer and a cyrillic locale (only players have seen this issue)
### Issue description
It seems that Godot 4 randomly fails to load `NodePath` export variables for... | bug,topic:core,needs testing,topic:dotnet | low | Critical |
2,348,611,516 | next.js | Not seeing not-found page on production deployment with dynamicParams = false | ### Link to the code that reproduces this issue
https://github.com/Cuboctaedro/nextjs-notfound
### To Reproduce
The app needs one env variable NEXT_PUBLIC_DOMAIN="http://localhost:3000/" on local or your domain if you deploy it.
Run on localhost and you will see urls like: /this-does-not-exist/page-d showing the no... | bug | low | Major |
2,348,615,636 | flutter | tool_crash ProcessException Process exited abnormally with exit code -10 Command xcrun OS error code -10 | flutter create .
Oops; flutter has exited unexpectedly: "ProcessException: Process exited abnormally with exit code -10:
Command: xcrun xcodebuild -list".
| c: crash,tool,P2,team-tool,triaged-tool | low | Critical |
2,348,624,573 | ollama | please add support for rk3588 NPU | Please add support for rk3588 NPU. Thanks. | feature request | low | Major |
2,348,658,502 | kubernetes | Optimize Pod informer memory efficiency used in admission plugins | ### What would you like to be added?
Memory optimizations on Pod informer used by admission plugins to reduce memory overhead of apiserver.
Currently, three admission plugins are using Pod informer:
1. ServiceAccount
2. NodeRestriction
3. PodSecurity
Optimization strategies:
For ServiceAccount & NodeRest... | priority/awaiting-more-evidence,sig/node,sig/api-machinery,kind/feature,help wanted,triage/accepted | low | Major |
2,348,723,239 | pytorch | torch.onnx.export - `repeat_interleave` produces invalid model | ### 🐛 Describe the bug
While adding ONNX export support for Cohere models to Optimum ([PR](https://github.com/huggingface/optimum/pull/1905)), I ran into an issue due to [a single problematic](https://github.com/huggingface/transformers/blob/254b25abd9a04bfff69408b9fdbaf2d270d1caaa/src/transformers/models/cohere/mo... | needs reproduction,module: onnx,triaged | low | Critical |
2,348,757,269 | ant-design | 获取 Tree 拖拽后的节点位置 dropPos 以更新 treeData | ### Reproduction link
[https://ant.design/components/tree-cn#components-tree-demo-draggable](https://ant.design/components/tree-cn#components-tree-demo-draggable)
### Steps to reproduce
拖动tree节点，触发 `onDrop` 事件
### What is expected?
在`onDrop`事件中得到拖动节点的位置信息`dragPos`和放置节点的位置信息`dropPos`，以便根据pos高效更新 `treeData... | Inactive,unconfirmed | low | Minor |
2,348,776,056 | stable-diffusion-webui | [Feature Request]: Stable Diffusion 3 Medium support | ### Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
### What would your feature do ?
Support Stable Diffusion 3 Medium
https://huggingface.co/stabilityai/stable-diffusion-3-medium
### Proposed workflow
Just like other SD models
### ... | enhancement | high | Critical |
2,348,803,134 | rust | Incorrect target feature shown during `asm!` verification (aarch64) | ### Code
```Rust
asm!("ldp d0, d1, [x0, 0x100]");
```
### Current output
```Shell
error: instruction requires: fp-armv8
--> /Users/lysan/.cargo/git/checkouts/unwinding-943b7a48f658f090/5718f01/src/unwinder/arch/aarch64.rs:101:13
|
101 | ldp d0, d1, [x0, 0x100]
| ^
... | A-diagnostics,T-compiler | low | Critical |
2,348,980,265 | pytorch | elements of STL get converted to an `IValue` and cannot be converted to an STL. | ### 🐛 Describe the bug
I think that the following snippet
```C++
#include <torch/extension.h>
struct h: torch::CustomClassHolder{
h(std::vector<std::pair<torch::ScalarType, std::vector<int64_t>>>){
}
void push(std::vector<std::pair<torch::ScalarType, std::vector<int64_t>>>){}
};
TORCH_LIBRAR... | module: cpp-extensions,module: cpp,triaged | low | Critical |
2,349,047,124 | flutter | [Web | CanvasKit | Chrome] Image with color representation of "Uncalibrated" does not appear | ### Steps to reproduce
1. Download the sample and run it on the Web (Chrome) with the default (canvas kit) renderer enabled
### Expected results
The vehicle image should appear successfully
### Actual results
A black rectangle appears instead
_Note: It works fine on Safari_
### Code sample
<details open><sum... | engine,platform-web,e: web_canvaskit,has reproducible steps,P3,browser: chrome-desktop,e: web_skwasm,team-web,triaged-web,found in release: 3.22,found in release: 3.23 | low | Major |
2,349,107,839 | stable-diffusion-webui | [Feature Request]: support simple crop anchor | ### Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
### What would your feature do ?
something like photoshop can choose anchor when resize crop
 sequence misbehaves when in "legacy ANSI" format | ### Windows Terminal version
N/A
### Windows build number
10.0.22631.0
### Other Software
neovim 0.10.0
zellij 0.40.1
wezterm 20240520-135708-b8f94c47
### Steps to reproduce
More details are described in https://github.com/wez/wezterm/issues/5450
### A short overview
In the wezterm issue mentioned above,... | Area-VT,Issue-Bug,Product-Terminal | medium | Critical |
2,349,181,460 | deno | Expose `zstd` compress to js code | `zstd` (crate) is a dependency of deno, but I want to use zstd compress in js code.
Currently it seems that there is no API for zstd, maybe add this through `DecompressionStream` ? | web,suggestion | low | Minor |
2,349,199,561 | transformers | LLM Cache | ### Feature request
Hi,
I'm working with an LLM model that generates text in stages. It stops after each section, performs an action, integrates the action's result into the generated text, and then continues. While this approach works, I'm facing a problem in time it takes.
I've tried using a static cache, but ... | Feature request | low | Minor |
2,349,291,354 | godot | Resource "Save as..." should change the resource path of current node, but now it change all the references | ### Tested versions
- reproduced in 4.2.2
### System information
macos
### Issue description
Resource "Save as..." should change the resource path of current node, but now it change all the references
### Steps to reproduce
1. create a AnimatedSprite and create SpriteFrames
2. save the SpriteFrames to local
3.... | discussion,topic:editor,needs testing | low | Minor |
2,349,329,193 | rust | Borrow checker error does not figure out source of `'static` requirement through closure capture | ### Code
```Rust
fn main() {
let targets = String::new();
let queue = targets.as_str();
(move || require_static(queue))();
}
fn require_static(_: &'static str) {}
```
### Current output
```Shell
error[E0597]: `targets` does not live long enough
--> src/main.rs:3:17
|
2 | let targets =... | A-diagnostics,P-medium,T-compiler,regression-from-stable-to-stable | low | Critical |
2,349,391,256 | flutter | Add an API for Getting Glyph Outlines as Paths | ### Use case
There are cases where apps need access to more than just the rendered text glyphs. This is especially common in design software where text glyphs need to be converted into shapes to allow for tweaking them, animating them, or styling them.
Skia for example has getPath:
https://github.com/google/skia... | c: new feature,framework,a: typography,c: proposal,P3,team-engine,triaged-engine | low | Minor |
2,349,427,742 | pytorch | test/dynamo/test_torchrec.py doesn't run in OSS CI (and won't run locally) | Putting up this issue. I was in the middle of fixing it and got distracted, I may or may not come back to it.
What's going on is:
- we run this test internally, which is why it hasn't bitrotten yet
- our fbgemm and torchrec commit pins need to be updated. (I can't seem to compile fbgemm from main in open source, t... | triaged,oncall: pt2,module: dynamo | low | Minor |
2,349,441,634 | storybook | [Bug]: infinite paths on Add a new story: ENAMETOOLONG. Function unavailable | ### Describe the bug
When attempting to document a new component, component search fails
BUG:
A large RED popup window with white text beginning
An error occurred while searching for components in the project. ENAMETOOLONG: name too long, scandir '/
IMPACT:
No components can be documented using the UI
may ... | bug,core,sev:S3 | low | Critical |
2,349,492,101 | rust | `async_closure`: error: implementation of `AsyncFnOnce` is not general enough | <!--
Thank you for filing a bug report! 🐛 Please provide a short summary of the bug,
along with any information you feel relevant to replicating the bug.
-->
I'm not sure if this is *really* a bug (or perhaps a duplicate issue of another bug). However, I can't see any reason why this shouldn't work.
I tried t... | T-compiler,C-bug,A-async-await,AsyncAwait-Triaged,T-types | low | Critical |
2,349,501,077 | vscode | Skip re-requesting implicit code actions if diagnostics around the cursor have not changed | https://github.com/microsoft/vscode/blob/a0185ee7cbcceda421361a2af44989dd40916a51/src/vs/editor/contrib/codeAction/browser/codeActionModel.ts#L57
I believe the implicit code actions are re-requested whenever diagnostics in the file change. Instead I proposed that we only re-request code actions if the diagnostics on ... | editor-code-actions,under-discussion | low | Minor |
2,349,501,943 | flutter | `Matrix4 entries must be finite. 'dart:ui/painting.dart'` | ### Steps to reproduce
use https://github.com/svthalia/Reaxit/tree/fe7c99c0852a7cbf8e44bbb73647ac473baa090d
Run the integration tests in github CI, try about 20 times and it should fail at least once. It is flaky
### Expected results
The tests should all pass
### Actual results
Sometimes the tests fail,... | a: tests,c: crash,platform-ios,framework,P3,team-framework,triaged-framework | low | Critical |
2,349,588,466 | pytorch | [torch.compile] Llama2 failure using dynamic shapes with Torch 2.4 nightly | ### 🐛 Describe the bug
torch.compile fails with the following error on Llama2. The error seems to occuring here : https://github.com/pytorch/TensorRT/blob/main/py/torch_tensorrt/dynamo/backend/backends.py#L90-L97
```py
RuntimeError: View operation returned a tensor that is the same as the input base tensor. This i... | triaged,oncall: pt2,module: dynamic shapes | low | Critical |
2,349,605,716 | godot | With Wayland enabled Godot windows often show the wrong icon in Alt+Tab switcher in GNOME | ### Tested versions
Reproducible: v4.3.beta1.official [a4f2ea91a]
Not Reproducible: 4.2.2-stable
### System information
Godot v4.3.beta1 - Fedora Linux 40 (Workstation Edition) - Wayland - Vulkan (Forward+) - dedicated NVIDIA GeForce RTX 3050 Laptop GPU () - AMD Ryzen 9 6900HS Creator Edition (16 Threads)
... | bug,platform:linuxbsd,topic:porting | low | Critical |
2,349,653,677 | flutter | On macOS Sequoia - [tool_crash] SocketException: Send failed, OS Error: No route to host, errno = 65 | ## Command
```sh
flutter run
```
## Steps to Reproduce
Run flutter app on IOS 18 Beta
## Logs
SocketException: Send failed, OS Error: No route to host, errno = 65
```console
#0 _NativeSocket.send (dart:io-patch/socket_patch.dart:1269:34)
#1 _RawDatagramSocket.send (dart:io-patch/socket_patch.dar... | c: crash,platform-ios,tool,customer: crowd,t: xcode,e: OS-version specific,P1,team-tool,triaged-tool | medium | Critical |
2,349,656,717 | bitcoin | docs: Wrong/outdated docs for `tr(KEY)` in doc/descriptors.md | Bitcoin Core uses BIP-386 for `tr(...)` descriptors, and in the case of `tr(KEY)`, the produced key is tweaked according to BIP-341:
https://github.com/bitcoin/bips/blob/85cda4e225b4d5fd7aff403f69d827f23f6afbbc/bip-0341.mediawiki?plain=1#L156
In the current descriptors.md, it is documented that `tr(KEY)` uses `K... | Docs,Descriptors | low | Minor |
2,349,666,967 | ollama | Add ollama run model --disable_model_load_on_start option. | I want an option that disables loading model on start of conversation. It will be helpful with another model requests from another client. | feature request | low | Minor |
2,349,679,913 | yt-dlp | southparkstudios.com is broken | ### DO NOT REMOVE OR SKIP THE ISSUE TEMPLATE
- [X] I understand that I will be **blocked** if I *intentionally* remove or skip any mandatory\* field
### Checklist
- [X] I'm reporting that yt-dlp is broken on a **supported** site
- [X] I've verified that I have **updated yt-dlp to nightly or master** ([update instruc... | site-bug | low | Critical |
2,349,686,449 | vscode | Cursor appears before decoration on empty lines | Type: <b>Bug</b>
Similar to or a regression of issue #40127
Start with decorator-sample example code and replace the body of updateDecorations with:
function updateDecorations() {
if (!activeEditor) {
return;
}
const testDecorationType = vscode.window.createTextEditorDecorationType({
before... | bug,editor-rendering | low | Critical |
2,349,687,437 | rust | rustdoc CURRENT_RUSTC_VERSION shows incorrect version between version bump and stage0 bump | Items marked with the stability attribute and `CURRENT_RUSTC_VERSION` will show the wrong version in nightly rustdoc output in between the version bump and the stage0 bump. That means in this sequence:
1. Item gets marked with `CURRENT_RUSTC_VERSION`. Example, on nightly the current rust version is 1.80, which is w... | T-rustdoc,T-bootstrap,A-docs,C-bug | low | Minor |
2,349,689,365 | TypeScript | 5.4.5 nuget package is missing. | ### Acknowledgement
- [X] I acknowledge that issues using this template may be closed without further explanation at the maintainer's discretion.
### Comment
[Nuget](https://www.nuget.org/packages/Microsoft.TypeScript.MSBuild#versions-body-tab) has 5.4.4 and 5.5.0.
| Visual Studio | low | Minor |
2,349,695,130 | flutter | Make it easier to develop for Android in the engine. | Today, it is possible to setup language servers for C++, Python, and Dart sources in the engine. However, I have not been able to find a reliable setup for editing Java code in the engine.
Anecdotally, editing java code is pretty miserable. I usually just have to resort to editing the file as plain text with no IDE... | P2,c: tech-debt,team-engine,triaged-engine | low | Minor |
2,349,696,194 | pytorch | aot_export_module produces unserializable GraphModules | ### 🐛 Describe the bug
GraphModules returned from aot_export_module cannot be saved and loaded successfully:
```
import torch
from torchvision.models import resnet18
def custom_compiler(model: torch.fx.GraphModule, example_inputs):
gm, _ = torch._functorch.aot_autograd.aot_export_module(
model,
... | triaged,oncall: pt2,oncall: export | low | Critical |
2,349,722,939 | pytorch | Dynamo graph breaks when passing a constant value to a nn module initialized within a function | ### 🐛 Describe the bug
```
import torch
class MyMod(torch.nn.Module):
def __init__(self, constant):
super().__init__()
self.register_buffer("buf", torch.randn([2,2]))
self.param = torch.nn.Parameter(torch.randn([2,2]))
self.constant = constant
def forward(self, x):
... | triaged,oncall: pt2 | low | Critical |
2,349,762,214 | flutter | [fuchsia] Flutter schedules too many fences when calling `Present` in the rasterized thread, causing a crash | ### Steps to reproduce
Sporadic.
### Expected results
A frame is rendered.
### Actual results
Crash. Stack trace below.
```
0x80006df14ef4 (libc.so - abort.cc:16) __llvm_libc::__abort_impl__
0x80006df06520 (libc.so - zx_panic_libc.cc:32) __zx_panic
0x534899ec (<_> - handle_close_many.cc:29) fidl::HLCPPOutgo... | c: crash,engine,platform-fuchsia,P3,team-engine,triaged-engine | low | Critical |
2,349,782,091 | flutter | Easier, more direct access to ink effects | ### Document Link
https://flutter.dev/go/layered-material-widgets
### What problem are you solving?
Ink effects are painted on an underlying `Material` widget, which has configuration options for elevation, border radius, clip behavior, and text style—even if none of those are necessary. Currently, the ability... | design doc,:scroll: | low | Minor |
2,349,807,918 | godot | Display DPI scaling (Microsoft Windows Setting) leads to incorrect canvas size in web export with ALL Canvas Resize Policy | ### Tested versions
- Reproductible in: 4.2.2.stable.official [15073afe3]
### System information
Godot v4.2.2.stable - Windows 10.0.22621 (Windows 11) - Vulkan (Forward+) - dedicated NVIDIA GeForce RTX 4070 (NVIDIA; 32.0.15.5585) - Intel(R) Core(TM) i7-9700K CPU @ 3.60GHz (8 Threads)
### Issue description
... | bug,platform:web,topic:porting | low | Critical |
2,349,811,726 | ollama | Suggestion for RFC7231 Compliant Endpoint for Model Deletion | **Description:**
Currently, the Ollama library’s DELETE /api/delete endpoint requires the model name to be provided in the request body. However, it would be beneficial and more in line with RFC7231 standards to support a URL-based model name specification. This approach is more intuitive and aligns with common RESTfu... | feature request,api | low | Minor |
2,349,818,305 | pytorch | internal asset failure | ### 🐛 Describe the bug
Traceback (most recent call last):
File "/home/ec2-user/experiments/kan/pykan/kan_exp_1.py", line 25, in <module>
model.train(dataset, opt="LBFGS", steps=50);
File "/home/ec2-user/experiments/kan/pykan/kan/KAN.py", line 898, in train
self.update_grid_from_samples(dataset['trai... | needs reproduction,triaged,module: linear algebra | low | Critical |
2,349,825,892 | vscode | Notebook Papercuts when Experimenting with LLMs | I’ve spent a significant amount of time in Jupyter Notebooks over the past few weeks, experimenting with language models (LLMs). My focus has been on testing various models, crafting prompts, and exploring different retrieval algorithms and prompts.
I’ve managed to get my work done effectively but also ran into foll... | under-discussion,notebook,papercut :drop_of_blood: | low | Minor |
2,349,852,823 | flutter | Engine repo: clang_format not inserting newline at EOF for C++/ObjC/shaders | We seem to be missing `InsertNewlineAtEOF: true` behavior, which Chromium style should include already. Tried briefly scattering it around our `.clang-format` files locally, but that didn't get it to budge. | engine,P2,team-engine,triaged-engine | low | Minor |
2,349,941,905 | godot | GltfDocument Memory Leak issue | ### Tested versions
reproduced in v4.0.4.stable
### System information
Android & iOS - Godot v4.0.4.stable - Mobile
### Issue description
```
func loadGltf():
var document = GLTFDocument.new()
var path = "res://Assets/demo.glbt"
var state = GLTFState.new()
document.append_from_file(path, state)
if is_in... | bug,needs testing,topic:import | low | Minor |
2,349,965,255 | godot | Array.remove( int position ) accepts String as an argument instead of throwing error | ### Tested versions
- Reproducible in: 3.5.3.stable, 3.5.stable, 3.3.4.stable
### System information
Apple M1 Pro 2021 running MacOS 14.1
### Issue description
In the [3.5 documentation](https://docs.godotengine.org/en/3.5/classes/class_array.html#class-array-method-remove), it states that the remove() m... | bug,topic:gdscript,confirmed | low | Critical |
2,349,994,044 | ant-design | Button组件默认样式 transition: all会导致额外的副作用 | ### What problem does this feature solve?
Button的transition: all会导致我在修改一些类似margin的属性时也会意外触发transition动画
### What does the proposed API look like?
建议Button里面需要的transition属性具体写出来，不用使用all
<!-- generated by ant-design-issue-helper. DO NOT REMOVE --> | Inactive | low | Major |
2,350,022,345 | rust | compiletest: Deficiencies and pitfalls of `//@ normalize-*` headers | These headers allow individual tests to define custom output-normalization rules, so that output snapshots are less sensitive to things like platform differences, line numbers, and timestamps.
Unfortunately, they have a few limitations and pitfalls:
- [x] The value parser for these headers is very lax, and when it ... | A-testsuite,T-bootstrap,E-medium,C-bug,A-compiletest | low | Minor |
2,350,045,696 | ant-design | TimePicker.RangePicker `format` should support array type | ```jsx
<TimePicker.RangePicker order={false} format={['当日 HH:mm', '次日 HH:mm']} />
``` | help wanted,💡 Feature Request,Inactive | medium | Major |
2,350,046,427 | rust | Several tests with `//@ needs-force-clang-based-tests` don't seem to run anywhere | (Follow-up to #126180 and #126257.)
There are currently 5 tests that use the `needs-force-clang-based-tests` header (formerly `needs-matching-clang`):
- `tests/run-make/cross-lang-lto-clang/Makefile` (this one does actually run)
- `tests/run-make/cross-lang-lto-pgo-smoketest/Makefile`
- `tests/run-make/cross-la... | A-testsuite,T-compiler,C-bug,A-run-make | low | Critical |
2,350,053,404 | PowerToys | Remapping Altgr to left Alt resulting in holding left Alt | ### Microsoft PowerToys version
0.81.0
### Installation method
PowerToys auto-update
### Running as admin
Yes
### Area(s) with issue?
Keyboard Manager
### Steps to reproduce
I tried to remap Altgr + arrow keys, which Windows interprets as left Ctrl + right Alt + arrow keys, into left Alt + arrow keys. However,... | Issue-Bug,Needs-Triage | low | Minor |
2,350,064,880 | vscode | Terminal Ignoring First Character Of Input Via `Terminal.sendText` | Type: <b>Bug</b>
So, I've been noticing this for a while now, but a lot of times (not _every_ time) when I am using the terminal in VSCode and start typing, the initial first key press doesn't get recognized. I just thought it was a quirk and is easy enough to ignore...
BUT...
I have an extension I have author... | bug,api,terminal-input | low | Critical |
2,350,067,082 | flutter | Text inside FittedBox is not properly aligned | ### Steps to reproduce
1. iOS 16，系统设置——> 显示——>缩放显示——>更大字体
2. Flutter 3.19版本开始，到最新版本Flutter3.22都存在这个问题
使用FittedBox进行屏幕适配，期望在不同尺寸的屏幕上展示的效果一致，发现了此问题；
**文本渲染的单词每一个字符都不在一条基准线上**
#### (English translation)
1. iOS 16, System Settings -> Display -> Zoom Display -> Larger Font
2. This problem exists from Flutter 3.... | framework,has reproducible steps,P2,team-framework,triaged-framework,found in release: 3.22,found in release: 3.23 | low | Major |
2,350,103,645 | godot | Incorrect lerp usage with deltatime for camera smoothing (and probably other things too) | ### Tested versions
- Issue was checked for and is present in all the recent major release branches (master, 3.x, 3.5, and 4.2)
### System information
N/A
### Issue description
The function that interpolates the camera position to the position it should be following isnt properly calculated with the frame delta, c... | bug,topic:rendering,needs testing,topic:2d | low | Major |
2,350,132,494 | godot | Incorrect flat shading | ### Tested versions
Godot 4.3 beta1
### System information
Godot v4.3.beta1 - Windows 10.0.19045 - Vulkan (Forward+) - dedicated Radeon RX 560 Series (Advanced Micro Devices, Inc.; 31.0.14001.45012) - Intel(R) Core(TM) i5-4570 CPU @ 3.20GHz (4 Threads)
### Issue description
I decided to test flat shading in Godot,... | enhancement,discussion,topic:rendering,topic:3d | low | Critical |
2,350,134,737 | flutter | [tool] PathAccessException when trying to delete stale assets in `flutter_assets` build directory | On stable 3.22.
Command: `flutter assemble`
```txt
PathAccessException: PathAccessException: Cannot delete file, path = 'C:\<redacted>\build\app\intermediates\flutter\debug\flutter_assets\assets/<redacted>.png' (OS Error: Access is denied.
, errno = 5)
at _File.throwIfError(file_impl.dart:675)
at _File._delet... | c: crash,P2,team-tool,triaged-tool | low | Critical |
2,350,148,180 | pytorch | TorchScript JIT UnicodeDecodeError: ‘utf-8’ codec can’t decode byte 0xe0 in position 0: unexpected end of data | ### 🐛 Describe the bug
Hi, I have the following Tokenizer class which I’m trying to jit to use in c++:
```
class Tokenizer(jit.ScriptModule):
def __init__(self):
super().__init__()
self.tokens_to_idx : Dict[str, int] = {...}
self.idx_to_tokens : Dict[int, str] = {...}
@jit.script_method
... | oncall: jit | low | Critical |
2,350,195,850 | pytorch | Lintrunner results inconsistent | https://hud.pytorch.org/hud/pytorch/pytorch/ade3d07483f609d802d2f567fa66ac0d5420254b/1?per_page=100
torch/_inductor/jagged_lowerings.py failure randomly disappears and re appears
https://hud.pytorch.org/hud/pytorch/pytorch/732b4e90740f5fe84d40f716f911c6dcbd9845fb/1?per_page=100
mod_tracker failure disappears a... | triaged,module: devx | low | Critical |
2,350,286,000 | ui | Input of type number is returning as string | ### Describe the bug
Utilizing the Form Component of Shadcn, I noticed that the Input Components of type number are returning me string values.
Now I have a schema here from ZOD
`likelihood_rating: z.number().int().min(1).max(5),`
And this is my formField:
```
<FormField
control={f... | bug | low | Critical |
2,350,293,228 | ollama | Models based on 'Qwen2ForCausalLM' are not yet supported | ### What is the issue?
I have downloaded the Qwen2-7b modle file(file type:safetensor), then using ollama create to load modle[ollama create QWEN2-7B -f qwen7b.modefile], the following is modle file content: FROM /app/antlaobr/models/Qwen2-7B
but get the follow error:
(base) antlabor@gpu001:/app/antlabor/models$ o... | bug | low | Critical |
2,350,348,424 | PowerToys | Functionality improvement of shortcut keys | ### Description of the new feature / enhancement
When pressing a shortcut key to call out a functional component, use "Shift+Win shift+V" to call the component of high level paste.
That is said you press Shift+Win , release pressing, and then press "shift+V".
The origin keys "Shift+Win+V" is hardly used for one h... | Needs-Triage | low | Minor |
2,350,355,496 | storybook | [Feature Request] Support ArgTypes inference from TS prop types in Svelte | ### Describe the bug
The Typescript types in a Svelte component don't come through to the props table in Storybook.
### Steps to reproduce the behaviour
`MyComponent.svelte`
```svelte
<script lang="ts">
/**
* Prop with type definition doesn't come through
*/
export let title: string;
/**
* P... | feature request,svelte,argtypes | low | Critical |
2,350,357,709 | rust | Strictly proof not tainted by errors in `TypeckResults` | I get here from #125888.
For now there is a hack when calling `tcx.typeck()`:
https://github.com/rust-lang/rust/blob/76c73827dcd0b363e60b22c3cef64bde4171bf17/compiler/rustc_hir_typeck/src/writeback.rs#L125-L130
so that in returned `TypeckResults`, the `tainted_by_errors` may be `None` even though there is actual... | I-ICE,T-compiler,T-types | low | Critical |
2,350,366,780 | tensorflow | "invalid static_cast" on AVX512FP16 (e.g. Sapphire Rapids) | ### Issue type
Bug
### Have you reproduced the bug with TensorFlow Nightly?
Yes
### Source
source
### TensorFlow version
2.15.1
### Custom code
No
### OS platform and distribution
RHEL 8
### Mobile device
_No response_
### Python version
3.10
### Bazel version
6.1.0
### GCC/compiler version
GCC 13.1
... | stat:awaiting tensorflower,type:bug,type:build/install,comp:core,TF 2.15 | low | Critical |
2,350,456,531 | pytorch | jacrev and jacfwd yield different results if one uses torch.no_grad blocks in module | ### 🐛 Describe the bug
When running a model through `torch.func.functional_call` that contains a `with torch.no_grad` block, you get different results when computing the jacobian with forward/reverse mode:
````python
class Model(nn.Module):
def __init__(self,):
super().__init__()
self.l1 ... | module: docs,module: autograd,triaged,actionable,module: functorch | low | Critical |
2,350,471,189 | deno | Bad Content-Length when use octokit upload asset with webStream | This code works fine
```ts
octokit.repos
.uploadReleaseAsset({
headers: {
"Content-Type": mime.getType(path.extname(filepath)) || "application/octet-stream",
"Content-Length": totalSize.toString(),
},
...baseParams,
release_id: release_id,
n... | bug,node compat | low | Critical |
2,350,483,827 | pytorch | Retry on CUBLAS_STATUS_ALLOC_FAILED | ### 🐛 Describe the bug
This is not technically a bug, so perhaps a feature request would be more appropriate.
We are encountering a scenario when running multiple libtorch threads we will periodically run out of memory which triggers the CUDA caching allocated to flush the free segments and retry the cudaMalloc. T... | module: cuda,triaged,enhancement,actionable,module: CUDACachingAllocator | low | Critical |
2,350,516,131 | ollama | Integration with MLFlow | Hey,
Currently, Ollama is saving models locally on a cache. To maintain different versions of LLMs or finetuned ones and also for extensive monitoring it's a good idea to provide integration with MLFlow where we can log all the experiments on MLFlow for better monitoring of the system.. I propose integrating Ollama ... | feature request | low | Minor |
2,350,528,999 | storybook | [Bug]: After pressing the "Reset" button, the object editor adds new object properties - ones that would give compilation errors when trying to change them using TS code | ### Describe the bug
TL;DR:
Intended behavior: I'm not allowed to change a property's value in my Angular/TS code - I get a compilation error ✔.
Then why does Storybook let me change that property's value? It shouldn't. ❌
---
Let's say I have a library of my own and I use the type **Config** from a 3rd party ... | bug,help wanted,addon: controls,sev:S3 | low | Critical |
2,350,571,552 | material-ui | [docs] Popover flip | ### Related page
https://mui.com/material-ui/react-popover/
### Kind of issue
Other
### Issue description
Hello, do MUI have a possibility to add modifiers to Popover component?
I wanna the same behaviour as Popper have, when dropdown have no space below the button it goes to top, and when no spacing top it goes ... | component: Popover,support: docs-feedback | low | Minor |
2,350,585,078 | ui | Select Items Text moves inwards | ### Describe the bug
When the Text Content of a SelectItem is longer it moves inwards. This only seems to happen if the Text is longer than the width of the Select component.
However for some SelectItems Texts
it moves inwards not at all and also the extent to which the text moves inwards varies.
### Affected component/components
Card
### How to reproduce
1. Go to https://u... | bug | low | Critical |
2,350,691,929 | ant-design | Progress组件circle圆形进度条支持起始位置设置 | ### What problem does this feature solve?
解决了圆形进度条，开始位置自定义，目前源码是12点位置开始，需要支持用户自定义
### What does the proposed API look like?
start-position
<!-- generated by ant-design-issue-helper. DO NOT REMOVE --> | 💡 Feature Request,Inactive | low | Minor |
2,350,766,501 | neovim | option scopes redesign | Currently options can have the following scopes:
- Buffer local. These options, despite the name, do still have a global value which is used as the default.
- Window local: Despite the name, these options can be dependent on the buffer as well. You can set a window option for all buffers on that window, or for a si... | refactor,compatibility,complexity:high,options,architecture | low | Critical |
Subsets and Splits
GitHub Issues Containing Next.js References
Filters training data to find examples mentioning "next.js", providing basic keyword search capability but offering limited analytical value beyond simple retrieval.