repo
stringclasses
11 values
fix_commit
stringlengths
40
40
buggy_commit
stringlengths
40
40
message
stringlengths
3
59.2k
files
listlengths
1
300
timestamp
timestamp[s]date
2016-04-11 04:39:55
2026-04-09 07:34:40
huggingface/transformers
931126b929ee573c2bb951db87bdbaea51567cea
c7064cdba11e9447fd1f64ec34e23c8b37113cb4
Fixes: Corrects file path for CUDA kernels (#37438) Corrects the file path used to locate the CUDA kernels for the Deformable Attention module. This ensures that the kernels are loaded correctly, resolving potential errors during module initialization and usage.
[ { "path": "src/transformers/models/deprecated/deta/modeling_deta.py", "patch": "@@ -57,7 +57,7 @@ def load_cuda_kernels():\n \n global MultiScaleDeformableAttention\n \n- root = Path(__file__).resolve().parent.parent.parent / \"kernels\" / \"deta\"\n+ root = Path(__file__).resolve().parent.parent....
2025-04-11T08:41:46
denoland/deno
486cb18fc54b408d40a02415e48e0cfb6cb196ed
f460188e583f00144000aa0d8ade08218d47c3c1
refactor(ext/webgpu): align error messages (#25719) Aligns the error messages in the ext/webgpu folder to be in-line with the Deno style guide. https://github.com/denoland/deno/issues/25269
[ { "path": "ext/webgpu/01_webgpu.js", "patch": "@@ -175,7 +175,7 @@ function assertDevice(self, prefix, context) {\n const deviceRid = device?.rid;\n if (deviceRid === undefined) {\n throw new DOMException(\n- `${prefix}: ${context} references an invalid or destroyed device.`,\n+ `${prefix}...
2024-09-19T07:14:54
huggingface/transformers
c7064cdba11e9447fd1f64ec34e23c8b37113cb4
371c44d0efb2ac552bbfed9a65baea4d4d83747c
enhance require_deterministic_for_xpu (#37437) * enhance require_deterministic_for_xpu Signed-off-by: YAO Matrix <matrix.yao@intel.com> * fix style Signed-off-by: YAO Matrix <matrix.yao@intel.com> * fix style Signed-off-by: YAO Matrix <matrix.yao@intel.com> --------- Signed-off-by: YAO Matrix <matrix.yao@intel....
[ { "path": "src/transformers/testing_utils.py", "patch": "@@ -139,7 +139,6 @@\n is_torch_available,\n is_torch_bf16_available_on_device,\n is_torch_bf16_gpu_available,\n- is_torch_deterministic,\n is_torch_fp16_available_on_device,\n is_torch_greater_or_equal,\n is_torch_hpu_availa...
2025-04-11T06:06:08
denoland/deno
f460188e583f00144000aa0d8ade08218d47c3c1
282c4c262d3a19fe3ae3fd9ea75811b816e65dc1
fix(ext/node): don't throw error for unsupported signal binding on windows (#25699)
[ { "path": "ext/node/polyfills/process.ts", "patch": "@@ -517,6 +517,12 @@ Process.prototype.on = function (\n // Ignores SIGBREAK if the platform is not windows.\n } else if (event === \"SIGTERM\" && Deno.build.os === \"windows\") {\n // Ignores SIGTERM on windows.\n+ } else if (\n+ ...
2024-09-19T03:22:01
huggingface/transformers
371c44d0efb2ac552bbfed9a65baea4d4d83747c
7ff896c0f22227ae999f991cff500e70452d5dcc
Remove old code for PyTorch, Accelerator and tokenizers (#37234) * Remove unneeded library version checks Signed-off-by: cyy <cyyever@outlook.com> * Remove PyTorch condition Signed-off-by: cyy <cyyever@outlook.com> * Remove PyTorch condition Signed-off-by: cyy <cyyever@outlook.com> * Fix ROCm get_device_capabi...
[ { "path": "src/transformers/activations.py", "patch": "@@ -16,7 +16,6 @@\n from collections import OrderedDict\n \n import torch\n-from packaging import version\n from torch import Tensor, nn\n \n from .utils import logging\n@@ -34,14 +33,6 @@ class PytorchGELUTanh(nn.Module):\n match due to rounding er...
2025-04-10T18:54:21
denoland/deno
282c4c262d3a19fe3ae3fd9ea75811b816e65dc1
fd23e8ec4f815f273a348d528edbc3bf65a21b5f
refactor(ext): align error messages (#25496) Aligns the error messages in the ext/http and a few messages in the ext/fetch folder to be in-line with the Deno style guide. This change-set also removes some unnecessary checks in the 00_serve.ts. These options were recently removed, so it doesn't make sense to check...
[ { "path": "ext/fetch/23_request.js", "patch": "@@ -122,7 +122,7 @@ function newInnerRequest(method, url, headerList, body, maybeBlob) {\n try {\n this.headerListInner = headerList();\n } catch {\n- throw new TypeError(\"cannot read headers: request closed\");\n+ t...
2024-09-19T01:19:45
huggingface/transformers
7d768764980ae093419d1b2bdc82bd1934b09bf2
dac443414ee99aeddf7c2a4b86ee5ff901cdd183
(Part 2) feat: allow for tp_size attr for tplizing the model (#37054) * feat: custom tp_size, new transformers tp interface Signed-off-by: Mehant Kammakomati <mehant.kammakomati2@ibm.com> * fix: review cmt - error when tp_plan not set for tp_size Signed-off-by: Mehant Kammakomati <mehant.kammakomati2@ibm.com> * fi...
[ { "path": "docs/source/ar/trainer.md", "patch": "@@ -674,29 +674,7 @@ use_cpu: false\n ```\n \n </hfoption>\n-<hfoption id=\"Tensor Parallelism with PyTorch 2\">\n \n-```yml\n-compute_environment: LOCAL_MACHINE\n-tp_config:\n- tp_size: 4\n-distributed_type: TP\n-downcast_bf16: 'no'\n-machine_rank: 0\n-main...
2025-04-10T15:44:09
denoland/deno
28dd92896251d02a912651aa56a2c81e47c107ee
315baeb50af0ff66961b36e0fcfc78f8dacb9198
fix: update nodeModulesDir config JSON schema (#25653) Ref https://github.com/denoland/deno/issues/25380
[ { "path": "cli/schemas/config-file.v1.json", "patch": "@@ -1,6 +1,6 @@\n {\n \"$id\": \"https://deno.land/x/deno/cli/schemas/config-file.v1.json\",\n- \"$schema\": \"http://json-schema.org/draft-07/schema\",\n+ \"$schema\": \"https://json-schema.org/draft/2020-12/schema\",\n \"description\": \"A JSON ...
2024-09-18T20:38:09
huggingface/transformers
dac443414ee99aeddf7c2a4b86ee5ff901cdd183
6daec12d0b2f922c8ad27b1cb0d51446d00d8016
fix: use mtime by default in Trainer._rotate_checkpoints with automatic fallback (#37260) Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
[ { "path": "src/transformers/trainer.py", "patch": "@@ -3224,9 +3224,8 @@ def _save_checkpoint(self, model, trial):\n \n # Maybe delete some older checkpoints.\n if self.args.should_save:\n- # Solely rely on numerical checkpoint id for rotation.\n- # mtime is not reliabl...
2025-04-10T15:42:06
denoland/deno
315baeb50af0ff66961b36e0fcfc78f8dacb9198
5057f58145ab47276fdc4392619888e37aea8864
ci: fix uploading canary (#25715) If a newer version was available the script was exiting with code 1.
[ { "path": ".github/workflows/ci.generate.ts", "patch": "@@ -818,6 +818,8 @@ const ci = {\n \"if [ $? -eq 0 ]; then\",\n \" echo ${{ github.sha }} > canary-latest.txt\",\n ' gsutil -h \"Cache-Control: no-cache\" cp canary-latest.txt gs://dl.deno.land/canary-$(rustc -vV |...
2024-09-18T20:17:12
huggingface/transformers
6daec12d0b2f922c8ad27b1cb0d51446d00d8016
0ea1151222b0ba4bf8e509e5e7ae73b57359d296
Add GGUF support to Gemma3 Text backbone (#37424) * add gemma3 gguf support Signed-off-by: Isotr0py <2037008807@qq.com> * fix typo and add gguf limit Signed-off-by: Isotr0py <2037008807@qq.com> * fix a typo Signed-off-by: Isotr0py <2037008807@qq.com> * add vision conversion test Signed-off-by: Isotr0py <2037008...
[ { "path": "src/transformers/integrations/ggml.py", "patch": "@@ -204,6 +204,23 @@\n \"attention.head_count\": \"num_attention_heads\",\n \"attention.head_count_kv\": \"num_key_value_heads\",\n \"attention.layer_norm_rms_epsilon\": \"rms_norm_eps\",\n+ \"attention.sliding_windo...
2025-04-10T15:15:43
denoland/deno
7a41a939972b701e96cb70cbf0516595fefcae02
ab1e391e1d700a68964e899963670e903f498cdf
fix(ext/http): gracefully handle Response.error responses (#25712) Fixes #14371
[ { "path": "ext/http/00_serve.ts", "patch": "@@ -511,9 +511,15 @@ function mapToCallback(context, callback, onError) {\n );\n }\n \n+ if (response.type === \"error\") {\n+ throw new TypeError(\n+ \"Return value from serve handler must not be an error response (like Response...
2024-09-18T19:14:40
huggingface/transformers
0ea1151222b0ba4bf8e509e5e7ae73b57359d296
9c0c323e12640a46090d2d13d135a1d84321feda
Llama Kernel integration (#37092) * initial commit * style * update * change approach attention * clean up * fix import * update * update * fix style * change method * attention * add mlp back * change name * update name * fix copies * fix config * fix
[ { "path": "src/transformers/integrations/hub_kernels.py", "patch": "@@ -31,7 +31,20 @@\n repo_id=\"kernels-community/deformable-detr\",\n layer_name=\"MultiScaleDeformableAttention\",\n )\n- }\n+ },\n+ \"RMSNorm\": {\n+ \"cuda\": La...
2025-04-10T15:13:25
denoland/deno
f347e779e0d1321d19a61de67295e7bd5edc3083
ec86089c4e95eabbf9edc48491031eb6711a410f
fix: error on unsupported compiler options (#25714) Update the config schema to error when adding unsupported `compilerOptions`. Fixes https://github.com/denoland/deno/issues/25696
[ { "path": "cli/schemas/config-file.v1.json", "patch": "@@ -9,6 +9,7 @@\n \"compilerOptions\": {\n \"type\": \"object\",\n \"description\": \"Instructs the TypeScript compiler how to compile .ts files.\",\n+ \"additionalProperties\": false,\n \"properties\": {\n \"allowJs\"...
2024-09-18T18:49:16
huggingface/transformers
9c0c323e12640a46090d2d13d135a1d84321feda
bde41d69b47c37e0dc1704cb4cd1a2a4709a4136
Fix require_read_token (#37422) * nit * fix * fix
[ { "path": "src/transformers/models/donut/modeling_donut_swin.py", "patch": "@@ -156,7 +156,7 @@ class DonutSwinImageClassifierOutput(ModelOutput):\n \"\"\"\n \n loss: Optional[torch.FloatTensor] = None\n- logits: torch.FloatTensor = None\n+ logits: Optional[torch.FloatTensor] = None\n hidd...
2025-04-10T15:01:40
denoland/deno
ec86089c4e95eabbf9edc48491031eb6711a410f
c90b07457958518a8830798d8800f647f897d12f
feat: require jsr prefix for `deno install` and `deno add` (#25698)
[ { "path": "cli/tools/registry/pm.rs", "patch": "@@ -437,18 +437,6 @@ pub async fn add(\n }\n \n let http_client = cli_factory.http_client_provider();\n-\n- let mut selected_packages = Vec::with_capacity(add_flags.packages.len());\n- let mut package_reqs = Vec::with_capacity(add_flags.packages.len());\...
2024-09-18T18:38:22
huggingface/transformers
bde41d69b47c37e0dc1704cb4cd1a2a4709a4136
7ecc5b88c0328aea91a3c9f8763f56b3b1e26767
Correctly drop tokens in SwitchTransformer (#37123) Previously, the identity function was used for dropped tokens with a weight from the expert that was not applied to the hidden states. This was misleading, because dropping means, the expert weight is zero. Instead of trying to fix the weight, we take an easier appro...
[ { "path": "src/transformers/models/switch_transformers/modeling_switch_transformers.py", "patch": "@@ -301,10 +301,8 @@ def forward(self, hidden_states):\n router_mask, router_probs, router_logits = self.router(hidden_states)\n expert_index = torch.argmax(router_mask, dim=-1)\n \n- # ...
2025-04-10T14:58:57
denoland/deno
c90b07457958518a8830798d8800f647f897d12f
49a0b7ab938145730354b7d6e575075b5e406622
fix(flags): don't treat empty run command as task subcommand (#25708)
[ { "path": "cli/args/flags.rs", "patch": "@@ -4679,7 +4679,7 @@ fn repl_parse(\n fn run_parse(\n flags: &mut Flags,\n matches: &mut ArgMatches,\n- app: Command,\n+ mut app: Command,\n bare: bool,\n ) -> clap::error::Result<()> {\n runtime_args_parse(flags, matches, true, true)?;\n@@ -4701,11 +4701,...
2024-09-18T18:36:37
huggingface/transformers
7ecc5b88c0328aea91a3c9f8763f56b3b1e26767
5ae9b2cac0b5b015e41eae328df56b97e8205262
Add image classifier donut & update loss calculation for all swins (#37224) * add classifier head to donut * add to transformers __init__ * add to auto model * fix typo * add loss for image classification * add checkpoint * remove no needed import * reoder import * format * consistency * add test of classif...
[ { "path": "docs/source/en/model_doc/donut.md", "patch": "@@ -226,3 +226,8 @@ print(answer)\n \n [[autodoc]] DonutSwinModel\n - forward\n+\n+## DonutSwinForImageClassification\n+\n+[[autodoc]] transformers.DonutSwinForImageClassification\n+ - forward\n\\ No newline at end of file", "additions": 5,...
2025-04-10T13:00:42
denoland/deno
49a0b7ab938145730354b7d6e575075b5e406622
3dd83fdbdba02bb59bc4af51f2ddcd8a85067398
fix: remove --allow-run warning when using deno without args or subcommand (#25684) Closes #25676 --------- Co-authored-by: David Sherret <dsherret@gmail.com>
[ { "path": "cli/args/flags.rs", "patch": "@@ -664,21 +664,33 @@ impl PermissionFlags {\n }\n \n pub fn to_options(&self) -> PermissionsOptions {\n+ fn handle_allow<T: Default>(\n+ allow_all: bool,\n+ value: Option<T>,\n+ ) -> Option<T> {\n+ if allow_all {\n+ assert!(value.is_n...
2024-09-18T15:05:11
huggingface/transformers
5ae9b2cac0b5b015e41eae328df56b97e8205262
d9e76656aea65c08e5cdd32470111f5883702155
Quark Quantization gated repo (#37412) * fix * empty commit * empty * nit * fix maybe ?
[ { "path": "tests/quantization/quark_integration/test_quark.py", "patch": "@@ -19,6 +19,7 @@\n is_torch_available,\n require_accelerate,\n require_quark,\n+ require_read_token,\n require_torch_gpu,\n require_torch_multi_gpu,\n slow,\n@@ -56,6 +57,7 @@ class QuarkTest(unittest.TestC...
2025-04-10T12:57:15
denoland/deno
bed46474b2d59b229bd85dbdec5b3d943c32f60f
48ea4e3c92b53936e89101a56a013300a47337d3
fix: do not panic running invalid file specifier (#25530) Co-authored-by: Bedis Nbiba <bedisnbiba@gmail.com>
[ { "path": "Cargo.lock", "patch": "@@ -1924,6 +1924,7 @@ dependencies = [\n \"libc\",\n \"log\",\n \"once_cell\",\n+ \"percent-encoding\",\n \"serde\",\n \"which 4.4.2\",\n \"winapi\",", "additions": 1, "deletions": 0 }, { "path": "cli/args/flags.rs", "patch": "@@ -34,6 +34,7 @@ use...
2024-09-18T13:51:39
huggingface/transformers
d9e76656aea65c08e5cdd32470111f5883702155
1ae8d54b0456867c83e28ecb1fde2e1dcd480e60
Fix new failure reports not including anything other than `tests/models/` (#37415) * fix * fix --------- Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
[ { "path": "utils/notification_service.py", "patch": "@@ -544,8 +544,11 @@ def payload(self) -> str:\n if \"https://github.com/huggingface/transformers/actions/runs\" in line:\n pattern = r\"<(https://github.com/huggingface/transformers/actions/runs/.+?/job/.+?)\\|(.+?)>\"...
2025-04-10T12:47:23
denoland/deno
7533492a7817ba61b3606dd412b78dc4d98faa76
42e3a49cc7d20ea8f9ba8984c93a47b5c7174f0f
tests: fix upgrade canary test (#25706)
[ { "path": "tests/specs/upgrade/canary/__test__.jsonc", "patch": "@@ -8,7 +8,7 @@\n },\n {\n \"commandName\": \"./deno_copy\",\n- \"args\": \"upgrade --canary\",\n+ \"args\": \"upgrade --canary --force\",\n \"output\": \"upgrade.out\",\n \"exitCode\": 0,\n \"flaky\":...
2024-09-18T13:43:50
huggingface/transformers
1ae8d54b0456867c83e28ecb1fde2e1dcd480e60
10144ff116497878d1d82aabf0a84c16944786f1
[chat-template] Unify tests and clean up 🧼 (#37275) * fix tests and some clean up * make one general test for each modality * remove redundant merging of kwargs * edge cases * dont enforce slow when reloading * fix gemma3 tests * has to adapt llama 4 after rebase * remove also from overriden tests * should b...
[ { "path": "docs/source/en/chat_templating_multimodal.md", "patch": "@@ -181,35 +181,6 @@ processed_chat = processor.apply_chat_template(\n print(processed_chat.keys())\n ```\n \n-</hfoption>\n-<hfoption id=\"custom frame sampling\">\n-\n-Some models don't sample frames *uniformly* and require more complex l...
2025-04-10T12:42:32
denoland/deno
f88a9e746744258ead3cc7a205644c18da5bab60
d5c00ef50e6519fccde54a577e038f0ebb1282e9
fix(dts): stabilize `fetch` declaration for use with `Deno.HttpClient` (#25683) This commit stabilizes the `fetch` function declaration for use with `Deno.HttpClient` and moves it from `lib.deno.unstable.d.ts` to `lib.deno.shared_globals.d.ts`. `Deno.HttpClient` was stabilized in #25569, but the associated overri...
[ { "path": "cli/tsc/dts/lib.deno.shared_globals.d.ts", "patch": "@@ -13,6 +13,7 @@\n /// <reference lib=\"deno.fetch\" />\n /// <reference lib=\"deno.websocket\" />\n /// <reference lib=\"deno.crypto\" />\n+/// <reference lib=\"deno.ns\" />\n \n /** @category WASM */\n declare namespace WebAssembly {\n@@ -74...
2024-09-18T06:38:21
huggingface/transformers
aa478567f824986e2ea0fab9739b409123824bef
ae5ce226644c8576c9047987e6b1d2e9bdeaed24
Allow rocm systems to run these tests (#37278) * Allow rocm systems to run these tests * Fix skipTest logic * Use get_device_properties to check system capabilities
[ { "path": "tests/models/musicgen/test_modeling_musicgen.py", "patch": "@@ -31,6 +31,7 @@\n T5Config,\n )\n from transformers.testing_utils import (\n+ get_device_properties,\n is_torch_available,\n require_flash_attn,\n require_torch,\n@@ -1093,12 +1094,15 @@ def test_sdpa_can_dispatch_on...
2025-04-10T11:33:01
denoland/deno
d5c00ef50e6519fccde54a577e038f0ebb1282e9
37315917625179063cb5653e2edd4ee0e5de99c5
feat(cli): evaluate code snippets in JSDoc and markdown (#25220) This commit lets `deno test --doc` command actually evaluate code snippets in JSDoc and markdown files. ## How it works 1. Extract code snippets from JSDoc or code fences 2. Convert them into pseudo files by wrapping them in `Deno.test(...)` 3. ...
[ { "path": "cli/args/flags.rs", "patch": "@@ -109,6 +109,8 @@ pub struct CacheFlags {\n #[derive(Clone, Debug, Eq, PartialEq)]\n pub struct CheckFlags {\n pub files: Vec<String>,\n+ pub doc: bool,\n+ pub doc_only: bool,\n }\n \n #[derive(Clone, Debug, Eq, PartialEq)]\n@@ -1694,6 +1696,19 @@ Unless --relo...
2024-09-18T04:35:48
huggingface/transformers
6d8b0b33786f6cf41791077624bf731f8c78e36b
f5865d32a2f56b3a83eb3349327abc2a9f16f3be
Fix Llama4 offset (#37414) * add +1 * Update modeling_llama4.py
[ { "path": "src/transformers/models/llama4/modeling_llama4.py", "patch": "@@ -731,7 +731,6 @@ def _update_causal_mask(\n attention_chunk_size = self.config.attention_chunk_size\n \n first_cache_position = cache_position[0]\n- last_cache_position = cache_position[-1]\n \n if pas...
2025-04-10T09:40:58
denoland/deno
97b8c9be3897f2a17439b217d1cd9884c39c087f
f360cae9dd1150c1ca02d51631ddaddb5bf871ab
refactor(ext/websocket): align error messages (#25622) Aligns the error messages in the ext/websocket folder to be in-line with the Deno style guide. https://github.com/denoland/deno/issues/25269
[ { "path": "ext/websocket/01_websocket.js", "patch": "@@ -161,14 +161,14 @@ class WebSocket extends EventTarget {\n \n if (wsURL.protocol !== \"ws:\" && wsURL.protocol !== \"wss:\") {\n throw new DOMException(\n- \"Only ws & wss schemes are allowed in a WebSocket URL.\",\n+ `Only ws &...
2024-09-17T22:32:12
huggingface/transformers
e032d12e8a402d3ec1ab7ec26b92d917f447af29
f834ca2c19215f1e4fb0959cc3faafeaf56cd4f7
the fix that did not get in (#37370) * debugging improvements * add debugging details * add more debugging details * debug more * the fix that did not get in * First fix flex * fix query offset * fix flex first * fix device mask creation for speed * small mask creation sdpa * Update flex_attention.py * remo...
[ { "path": "src/transformers/cache_utils.py", "patch": "@@ -1913,37 +1913,38 @@ def initialise_cache_layer(self, layer_idx, key_states):\n self.value_cache.append(new_layer_value_cache)\n \n def _sliding_update(self, cache_position, layer_idx, key_states, value_states, k_out, v_out, max_cache_len...
2025-04-09T18:15:33
denoland/deno
f360cae9dd1150c1ca02d51631ddaddb5bf871ab
d4a06251c54fc004e189469e493b1261be200300
fix(lsp): properly resolve jsxImportSource for caching (#25688)
[ { "path": "cli/lsp/language_server.rs", "patch": "@@ -1,6 +1,5 @@\n // Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.\n \n-use base64::Engine;\n use deno_ast::MediaType;\n use deno_config::workspace::WorkspaceDirectory;\n use deno_config::workspace::WorkspaceDiscoverOptions;\n@@ -96...
2024-09-17T17:29:19
huggingface/transformers
f834ca2c19215f1e4fb0959cc3faafeaf56cd4f7
c5c648dd74b711992a6af114c3b1524aece99b97
Attention Quantization with FBGemm & TP (#37384) * fix * keep fused * contiguous * rm print * update * update * rm print
[ { "path": "src/transformers/integrations/fbgemm_fp8.py", "patch": "@@ -50,7 +50,7 @@ def forward(self, x):\n # x_quantized and x_scale are not necessarily on the same device as x, this is an issue.\n # https://github.com/pytorch/FBGEMM/blob/e08af8539c391437f447173863df0f3f6f6f1855/fbgemm_gpu...
2025-04-09T16:45:42
huggingface/transformers
c5c648dd74b711992a6af114c3b1524aece99b97
71b35387fd6d71487bd29e694ed10d925203e031
Fix some failing AWQ tests (#37383) * update AwqQuantizer * fix style * add an arg to get_modules_to_not_convert to add get_keys_to_not_convert(model)
[ { "path": "src/transformers/quantizers/base.py", "patch": "@@ -263,14 +263,17 @@ def get_modules_to_not_convert(\n model: \"PreTrainedModel\",\n skip_modules: Optional[List[str]] = None,\n keep_in_fp32_modules: Optional[List[str]] = None,\n+ add_default_skips: bool = False,\n ...
2025-04-09T16:24:57
denoland/deno
063f427ddf52386b006db9295d534578cf60412b
a14e9f55b4efc39689e1cc76b44bc6c1be51efa5
fix(ext/node): stub `inspector/promises` (#25635) Signed-off-by: Divy Srivastava <dj.srivastava23@gmail.com>
[ { "path": "ext/node/lib.rs", "patch": "@@ -615,6 +615,7 @@ deno_core::extension!(deno_node,\n \"node:http2\" = \"http2.ts\",\n \"node:https\" = \"https.ts\",\n \"node:inspector\" = \"inspector.ts\",\n+ \"node:inspector/promises\" = \"inspector.ts\",\n \"node:module\" = \"01_require.js\",\...
2024-09-17T17:27:19
huggingface/transformers
71b35387fd6d71487bd29e694ed10d925203e031
ad340908e441246f59462ee4f3450085569e4f8f
Apply torchfix to replace deprecated functions: `_pytree._register_pytree_node` and `torch.cpu.amp.autocast` (#37372) fix: apply torchfix
[ { "path": "src/transformers/trainer.py", "patch": "@@ -3689,7 +3689,7 @@ def autocast_smart_context_manager(self, cache_enabled: Optional[bool] = True):\n arguments, depending on the situation.\n \"\"\"\n if self.use_cpu_amp:\n- ctx_manager = torch.cpu.amp.autocast(cache_e...
2025-04-09T15:11:18
denoland/deno
915b73571c40601b7f79fa4ee0f3035982e23b7f
039e15fc84465fd2595d4fc74bf817a44ba371ec
ci: fix releasing canary (#25689)
[ { "path": ".github/workflows/ci.generate.ts", "patch": "@@ -813,8 +813,8 @@ const ci = {\n ].join(\"\\n\"),\n run: [\n 'gsutil -h \"Cache-Control: public, max-age=3600\" cp ./target/release/*.zip gs://dl.deno.land/canary/$(git rev-parse HEAD)/',\n- 'lastest_canary_...
2024-09-17T16:53:32
huggingface/transformers
ad340908e441246f59462ee4f3450085569e4f8f
2527f71a47b267f2ea7f4afc71a340106dd08809
Fix warning message for PEFT models in text-generation pipeline #36783 (#36887) * add peft model in constant * add test * fix formating * make fixup execute * change code * check by self.task * add test * fixup test code * fix minor typo * fix pipeline test * apply maintainers reqests
[ { "path": "src/transformers/pipelines/base.py", "patch": "@@ -849,6 +849,22 @@ def build_pipeline_init_args(\n supports_binary_output=True,\n )\n \n+SUPPORTED_PEFT_TASKS = {\n+ \"document-question-answering\": [\"PeftModelForQuestionAnswering\"],\n+ \"feature-extraction\": [\"PeftModelForFeatureEx...
2025-04-09T14:36:52
denoland/deno
f7ddea3af7a9f4dfef23aa544f05348dabbad20d
51d926ac30413a6747505b3f248f43a563470545
chore: lint `40_fs_events.js` (#25672) Fixes CI https://github.com/denoland/deno/actions/runs/10892648144/job/30225971485
[ { "path": "runtime/js/40_fs_events.js", "patch": "@@ -1,6 +1,6 @@\n // Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.\n \n-import { core, internals, primordials } from \"ext:core/mod.js\";\n+import { core, primordials } from \"ext:core/mod.js\";\n import { op_fs_events_open, op_fs_e...
2024-09-16T22:54:57
huggingface/transformers
7ae0be722edf891f7686fc79d1d8cb12ccc44bc4
e3eda6d18864045c2e8a1884914dbeda4998acbe
update deepspeed docker (#37371) * update * create docker image * 03 * uninstall pytest as it conflits with transformers * wrong one * better * see which package depends on pytest * up * resintall * fix * deepspeedddddddd * deepspeedddddddd * deepspeedddddddd * deepspeedddddddd * deepspeedddddddd * deep...
[ { "path": ".github/workflows/build-docker-images.yml", "patch": "@@ -63,14 +63,14 @@ jobs:\n uses: huggingface/hf-workflows/.github/actions/post-slack@main\n with:\n slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER }}\n- title: 🤗 Results of the transformers-all-latest-g...
2025-04-09T12:54:06
denoland/deno
8fe90d286160d6cc585e18b1a2f4e29bd1523852
e81c67ba227b00ccde495fb0afb6ae14f69a3a30
fix(npm): better error handling for remote npm deps (#25670) * https://github.com/denoland/deno_npm/pull/68
[ { "path": "Cargo.lock", "patch": "@@ -1868,9 +1868,9 @@ dependencies = [\n \n [[package]]\n name = \"deno_npm\"\n-version = \"0.25.0\"\n+version = \"0.25.1\"\n source = \"registry+https://github.com/rust-lang/crates.io-index\"\n-checksum = \"efb2dd7efaf478d780a6ca43e6127667d5329ed9a5ff89f42cec0dc21ee58342\"...
2024-09-16T21:23:40
huggingface/transformers
e3eda6d18864045c2e8a1884914dbeda4998acbe
1e6ff5fd5538d81e653d3752284e725a04ff0a26
Add glm4 (#37388) * add changed * Revert "add changed" This reverts commit 0a0166a1fe80556115a49fbf0c2132de0f4f85c9. * update with NEW MODEL class called GLM4 * update * Update glm4.md * Name * style * fix copies * fixup test --------- Co-authored-by: Yuxuan Zhang <2448370773@qq.com>
[ { "path": "docs/source/en/_toctree.yml", "patch": "@@ -461,6 +461,8 @@\n title: Gemma2\n - local: model_doc/glm\n title: GLM\n+ - local: model_doc/glm4\n+ title: glm4\n - local: model_doc/openai-gpt\n title: GPT\n - local: model_doc/gpt_neo", "additi...
2025-04-09T12:02:04
denoland/deno
6ce16145dd12d8a272cb543871276c33c8201a37
36a1a79f175b2df31d8ba5d7e121239c3b7e74ea
chore: maybe fix flaky serve_parallel test (#25668)
[ { "path": "tests/integration/serve_tests.rs", "patch": "@@ -39,15 +39,18 @@ impl ServeClientBuilder {\n \n ServeClient::with_child(child)\n }\n+\n fn map(\n self,\n f: impl FnOnce(util::TestCommandBuilder) -> util::TestCommandBuilder,\n ) -> Self {\n Self(f(self.0), self.1)\n }\n+\n ...
2024-09-16T18:55:49
huggingface/transformers
1e6ff5fd5538d81e653d3752284e725a04ff0a26
6f4058aee3dd2b3b4cba290a008692e399efbd72
fix: llama4 conversion script no_rope_layers (#37359) fix conversion script no_rope_layers `no_rope_layers` should either be a list of NoPE layers or None, such that it is created in the config from the `no_rope_layer_interval` Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
[ { "path": "src/transformers/models/llama4/convert_llama4_weights_to_hf.py", "patch": "@@ -286,7 +286,6 @@ def write_model(\n tie_word_embeddings=False, # Constant set to False\n torch_dtype=torch_dtype,\n for_llm_compressor=_OFFLINE_QUANT_COMPATIBLE,\n- no_rope_layers=no_rope...
2025-04-09T11:02:15
denoland/deno
e47606a5549aff43540bf2b8adc5db0309c28dbc
968f441da32ae5fc3783a4f9bcda3d34fa276b4d
fix(lockfile): use loose deserialization for version constraints (#25660) * https://github.com/denoland/deno_lockfile/pull/39 Closes https://github.com/denoland/deno/issues/25649
[ { "path": "Cargo.lock", "patch": "@@ -1719,9 +1719,9 @@ dependencies = [\n \n [[package]]\n name = \"deno_lockfile\"\n-version = \"0.23.0\"\n+version = \"0.23.1\"\n source = \"registry+https://github.com/rust-lang/crates.io-index\"\n-checksum = \"fb68a4a666c69eabd8fe505d6fdc4ed4a2d962fe1680dbfa8b0c8a2975d58...
2024-09-16T15:29:21
denoland/deno
8fa92228bb748bdc59e0fe003108dcaea0a18d10
74069add3f400ad6b232ad18840206ab96b9fe80
fix(types): simplify mtls related types (#25658) Instead of two overloads for `Deno.connectTls` and `Deno.createHttpClient`, there is now just one.
[ { "path": "cli/tsc/dts/lib.deno.ns.d.ts", "patch": "@@ -6107,7 +6107,12 @@ declare namespace Deno {\n \n /** Create a custom HttpClient to use with {@linkcode fetch}. This is an\n * extension of the web platform Fetch API which allows Deno to use custom\n- * TLS certificates and connect via a proxy w...
2024-09-16T12:35:55
huggingface/transformers
6f4058aee3dd2b3b4cba290a008692e399efbd72
08e3217bafddc5d11ce0e7369bcfaaabe5501ba5
Update composition flag usage (#36263) * update composition flag usage * remove print * fix tests * actually fix * oh c'mon * now should be fixed right? * fix copies
[ { "path": "src/transformers/cache_utils.py", "patch": "@@ -1195,9 +1195,7 @@ def __init__(\n self.max_cache_len = config.max_position_embeddings if max_cache_len is None else max_cache_len\n \n # Some model define a custom `head_dim` != config.hidden_size // config.num_attention_heads\n- ...
2025-04-09T09:48:49
huggingface/transformers
4d0de5f73a0715ac9a63de5fa461af48d3abbb10
c15a7adb283fa984a40558c7fe7bed30ae975cdd
:rotating_light: :rotating_light: Setup -> setupclass conversion (#37282) * More limited setup -> setupclass conversion * make fixup * Trigger tests * Fixup UDOP * Missed a spot * tearDown -> tearDownClass where appropriate * Couple more class fixes * Fixups for UDOP and VisionTextDualEncoder * Ignore errors w...
[ { "path": "tests/models/altclip/test_processor_altclip.py", "patch": "@@ -26,15 +26,16 @@\n class AltClipProcessorTest(ProcessorTesterMixin, unittest.TestCase):\n processor_class = AltCLIPProcessor\n \n- def setUp(self):\n- self.model_id = \"BAAI/AltCLIP\"\n- self.tmpdirname = tempfile....
2025-04-08T16:15:37
huggingface/transformers
c15a7adb283fa984a40558c7fe7bed30ae975cdd
121f91d36c171b67c62320507dfaa460eab7657c
fix(qwen): fix shape error when using tp (#36947) * fix(qwen): fix shape error when using tp * Update modeling_qwen2_vl.py --------- Co-authored-by: shidongxing <shidongxing@pjlab.org.cn> Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
[ { "path": "src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py", "patch": "@@ -961,7 +961,7 @@ def forward(\n )\n \n attn_output = attn_output.transpose(1, 2).contiguous()\n- attn_output = attn_output.view(bsz, q_len, self.hidden_size)\n+ attn_output = attn_output.view(bs...
2025-04-08T15:47:30
denoland/deno
74069add3f400ad6b232ad18840206ab96b9fe80
f8547e2617a5bbef4931ac0df6e70db6b901927f
fix(runtime): don't error `child.output()` on consumed stream (#25657) This fixes the fast path for `readableStreamCollectIntoUint8Array` to only trigger if the readable stream has not yet been disturbed - because otherwise we may not be able to close it if the read errors.
[ { "path": "ext/web/06_streams.js", "patch": "@@ -1054,7 +1054,7 @@ async function readableStreamCollectIntoUint8Array(stream) {\n getReadableStreamResourceBackingUnrefable(stream);\n const reader = acquireReadableStreamDefaultReader(stream);\n \n- if (resourceBacking) {\n+ if (resourceBacking && !is...
2024-09-16T12:23:40
huggingface/transformers
121f91d36c171b67c62320507dfaa460eab7657c
4321b0648c4d80f8d0e3b02d555521c7e1483f46
prune LM Head for USD (#36695) * initial commit * fix * fix style * set default to prune * add tests * comment * remove prune flag from generate * address Joao's comments * deprecate_kwarg * add doc * fix target_vocab_size * Update src/transformers/generation/candidate_generator.py Co-authored-by: Joao Gan...
[ { "path": "src/transformers/generation/candidate_generator.py", "patch": "@@ -19,7 +19,9 @@\n \n import numpy as np\n import torch\n+import torch.nn as nn\n \n+from ..pytorch_utils import prune_linear_layer\n from ..utils import is_sklearn_available\n \n \n@@ -36,6 +38,8 @@\n from ..tokenization_utils_b...
2025-04-08T15:44:10
denoland/deno
81c9e0ba25acb6288330b85ba2c558f587d92782
eb8ee95f08186c948e5b83501cedd59d6e3b4ef2
fix(ext/crypto): support md4 digest algorithm (#25656) Fixes #25646
[ { "path": "ext/node/ops/crypto/digest.rs", "patch": "@@ -80,6 +80,10 @@ macro_rules! match_fixed_digest_with_eager_block_buffer {\n type $type = ::sm3::Sm3;\n $body\n }\n+ \"rsa-md4\" | \"md4\" | \"md4withrsaencryption\" => {\n+ type $type = ::md4::Md4;\n+ $body\n+ ...
2024-09-16T11:04:40
huggingface/transformers
aab087832724849463cf5b28a40a82ea97777b14
35f0f5b5daea433ed2587aa9f72781a8de0c4448
Skip non-selected experts for mixtral and qwen2_moe (#32429) * Skip non-selected experts for mixtral and qwen2_moe * Fix: tensor tolist() * WIP: tokenization test * fix modular source of truth * nits --------- Co-authored-by: Arthur Zucker <arthur.zucker@gmail.com> Co-authored-by: Arthur <48595927+ArthurZucker@u...
[ { "path": "src/transformers/models/mixtral/modeling_mixtral.py", "patch": "@@ -135,11 +135,10 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:\n # this will be used to easily index which expert is going to be sollicitated\n expert_mask = torch.nn.functional.one_hot(selected...
2025-04-08T15:41:28
denoland/deno
a666c8c9f92d87e7a3b4d4e06fdc027b3bf9663e
db6fc12b9e2d5cb55c1f3c61b054805505ae11a6
fix(cli): `deno task` exit with status 0 (#25637) Fixes https://github.com/denoland/deno/issues/25632 Exit code 1 indiciates some sort of failure but `deno task` (without arguments) is used to list available commands. --------- Co-authored-by: Yoshiya Hinosawa <stibium121@gmail.com>
[ { "path": "cli/tools/task.rs", "patch": "@@ -75,7 +75,7 @@ See https://docs.deno.com/go/config\"#\n &cli_options.start_dir,\n &tasks_config,\n )?;\n- return Ok(1);\n+ return Ok(0);\n }\n };\n ", "additions": 1, "deletions": 1 }, { "path": "tests/specs/ta...
2024-09-15T05:59:24
huggingface/transformers
8064cd9b4fdd797843c2d884c309a3f2a35a5b25
cdfb018d0300fef3b07d9220f3efe9c2a9974662
fix deepspeed job (#37284) Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
[ { "path": ".github/workflows/self-scheduled.yml", "patch": "@@ -336,6 +336,10 @@ jobs:\n working-directory: ${{ inputs.working-directory-prefix }}/transformers\n run: git fetch && git checkout ${{ github.sha }}\n \n+ # TODO: update the docker image instead\n+ - name: Reinstall some...
2025-04-08T13:19:33
denoland/deno
db6fc12b9e2d5cb55c1f3c61b054805505ae11a6
4655172c0bfe6bd4f22b9e53b63bf1f0533f36cc
fix(ext/node): add `vm.constants` (#25630)
[ { "path": "ext/node/polyfills/vm.js", "patch": "@@ -27,7 +27,7 @@ import { ERR_INVALID_ARG_TYPE } from \"ext:deno_node/internal/errors.ts\";\n \n import { primordials } from \"ext:core/mod.js\";\n \n-const { Symbol, ArrayPrototypeForEach } = primordials;\n+const { Symbol, ArrayPrototypeForEach, ObjectFreeze...
2024-09-15T02:45:28
huggingface/transformers
2da82e432dbc08f9e497b353cdccfee7e84bd6a8
794fde7b1c3d041519fc28ea3e1461b0cfcad4e7
Multiple llama4 fixe (#37353) * update for fixes * more fixes * fuxix dynamic cache? * style * fix both traiining and generating. Eager seems alright * dynamic does not work * fix most cases, use_cache or not, eager or not, no default cache (ex: not training but you want to get cache states) * should be final f...
[ { "path": "src/transformers/cache_utils.py", "patch": "@@ -1857,7 +1857,7 @@ class HybridChunkedCache(Cache):\n \n # TODO (joao): dive deeper into gemma2 and paligemma -- there are reports of speed loss with compilation. Revert\n # ALL changes from the PR that commented the line below when reactivat...
2025-04-08T09:14:49
denoland/deno
4655172c0bfe6bd4f22b9e53b63bf1f0533f36cc
ccd1ca8a8b36bfd6640e9a47df498139a0b3d6cc
fix(ext/node): export `process.allowedNodeEnvironmentFlags` (#25629)
[ { "path": "ext/node/polyfills/process.ts", "patch": "@@ -748,6 +748,8 @@ Object.defineProperty(Process.prototype, \"allowedNodeEnvironmentFlags\", {\n },\n });\n \n+export const allowedNodeEnvironmentFlags = ALLOWED_FLAGS;\n+\n Process.prototype.features = { inspector: false };\n \n // TODO(kt3k): Get the...
2024-09-15T02:45:19
huggingface/transformers
794fde7b1c3d041519fc28ea3e1461b0cfcad4e7
b54c2f46891149210dbbe118fca55b1357a47003
Fixing flex attention for torch=2.6.0 (#37285) * adding compile kwarg for torch 2.6 * fixing dynamic * addressing comment * typo * Update src/transformers/integrations/flex_attention.py --------- Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
[ { "path": "src/transformers/integrations/flex_attention.py", "patch": "@@ -31,6 +31,7 @@\n import torch\n \n from ..utils import is_torch_flex_attn_available\n+from ..utils.import_utils import _torch_version\n \n \n if is_torch_flex_attn_available():\n@@ -60,8 +61,16 @@ def __init__(self):\n \"\"\"\...
2025-04-07T21:04:46
denoland/deno
ccd1ca8a8b36bfd6640e9a47df498139a0b3d6cc
597f2d8d4d9b91ee586b9787d6ba52d247e4ff87
fix(ext/node): add stubs for `node:trace_events` (#25628)
[ { "path": "ext/node/lib.rs", "patch": "@@ -621,6 +621,7 @@ deno_core::extension!(deno_node,\n \"node:timers\" = \"timers.ts\",\n \"node:timers/promises\" = \"timers/promises.ts\",\n \"node:tls\" = \"tls.ts\",\n+ \"node:trace_events\" = \"trace_events.ts\",\n \"node:tty\" = \"tty.js\",\n ...
2024-09-15T02:45:09
huggingface/transformers
b54c2f46891149210dbbe118fca55b1357a47003
754a370bca8f2733c36deacf92f5709310477016
more fixes for post-training llama4 (#37329) * more fixes for post-training llama4 * use target_length instead of guearded past_key_values
[ { "path": "src/transformers/models/llama4/modeling_llama4.py", "patch": "@@ -730,6 +730,7 @@ def forward(\n )\n return output if return_dict else output.to_tuple()\n \n+ @torch.compiler.disable # the operations in this method are not compilable\n def _update_causal_mask(\n se...
2025-04-07T19:20:23
denoland/deno
597f2d8d4d9b91ee586b9787d6ba52d247e4ff87
af2d992ecd2b9320072164b6ee295c31a3194406
feat: print `Listening on` messages on stderr instead of stdout (#25491) Fixes https://github.com/denoland/deno/issues/25114 --------- Signed-off-by: Leo Kettmeir <crowlkats@toaxl.com> Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com> Co-authored-by: crowlkats <crowlkats@toaxl.com> Co-authored-by: Nathan ...
[ { "path": "ext/http/00_serve.ts", "patch": "@@ -646,7 +646,7 @@ function serve(arg1, arg2) {\n options.onListen(listener.addr);\n } else {\n // deno-lint-ignore no-console\n- console.log(`Listening on ${path}`);\n+ console.error(`Listening on ${path}`);\n }\n })...
2024-09-14T21:30:06
denoland/deno
2c0bf6fd65b3276c4818c73028b19819608c8e70
d162733db2944b4c35d8c3e80f37f7de843dafa1
fix(ext/node): attach console stream properties (#25617) `kBindStreamsLazy` should be called with `process` during init, but it never was.
[ { "path": "ext/node/polyfills/01_require.js", "patch": "@@ -128,6 +128,7 @@ import internalTestBinding from \"ext:deno_node/internal/test/binding.ts\";\n import internalTimers from \"ext:deno_node/internal/timers.mjs\";\n import internalUtil from \"ext:deno_node/internal/util.mjs\";\n import internalUtilIns...
2024-09-14T00:28:35
denoland/deno
d162733db2944b4c35d8c3e80f37f7de843dafa1
8539cacbbe7dd7e5bece23b971cf1792f9ade3e7
fix(ext/node): use primordials in ext/node/polyfills/wasi.ts (#25608) Toward #24236
[ { "path": "ext/node/polyfills/wasi.ts", "patch": "@@ -1,7 +1,8 @@\n // Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.\n \n-// TODO(petamoriken): enable prefer-primordials for node polyfills\n-// deno-lint-ignore-file prefer-primordials\n+import { primordials } from \"ext:core/mod.js...
2024-09-13T17:01:07
denoland/deno
1270d9bc93dc6bceb0ae45b73ca154bbcf1a5db8
f2b53d42acf96044e56b21808cfc3f33a9501ec3
fix(check): move is cjs check from resolving to loading (#25597) This is required to do when loading because TypeScript handles and resolves `/// <reference path="..." />` internally.
[ { "path": "cli/tsc/99_main_compiler.js", "patch": "@@ -117,27 +117,6 @@ delete Object.prototype.__proto__;\n }\n }\n \n- class SpecifierIsCjsCache {\n- /** @type {Set<string>} */\n- #cache = new Set();\n-\n- /** @param {[string, ts.Extension]} param */\n- maybeAdd([specifier, ext]) {\n- ...
2024-09-13T13:36:26
huggingface/transformers
e2b0224d9489f92b7e4e2d898a53e209d977494c
6cc109c354ec9e9e306a3e7e5a212a8c86030c17
Update Model Card for Jamba (#37152) * Update model card for jamba * Apply the suggestions from code review Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Apply suggestions from code review-2 Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * update model page. *...
[ { "path": "docs/source/en/model_doc/jamba.md", "patch": "@@ -14,96 +14,126 @@ rendered properly in your Markdown viewer.\n \n -->\n \n+<div style=\"float: right;\">\n+ <div class=\"flex flex-wrap space-x-1\">\n+ <img alt=\"PyTorch\" src=\"https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytor...
2025-04-07T18:02:59
denoland/deno
606b7b17c6b5eccae9950eeb11d30e63e544b49d
7477c2d70639962a40c7333e766b4a4b4aa75ddd
refactor(runtime): align error messages (#25563) Aligns the error messages in the runtime folder to be in-line with the Deno style guide. https://github.com/denoland/deno/issues/25269
[ { "path": "runtime/js/10_permissions.js", "patch": "@@ -95,7 +95,7 @@ class PermissionStatus extends EventTarget {\n */\n constructor(status = null, key = null) {\n if (key != illegalConstructorKey) {\n- throw new TypeError(\"Illegal constructor.\");\n+ throw new TypeError(\"Illegal const...
2024-09-13T09:38:45
huggingface/transformers
6cc109c354ec9e9e306a3e7e5a212a8c86030c17
8bbcdf54096efbbb8de6f1a4e73c6517edf344a8
Improvements in Gemma2 model card (#37076) * Improved Model card for Gemma2 * Made changes in gemma2 as suggested * Made more changes in the doc (adding image, notes, closing hfoptions) * minor fixes --------- Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>
[ { "path": "docs/source/en/model_doc/gemma2.md", "patch": "@@ -14,36 +14,133 @@ specific language governing permissions and limitations under the License.\n rendered properly in your Markdown viewer.\n \n -->\n+<div style=\"float: right;\">\n+ <div class=\"flex flex-wrap space-x-1\">\n+ <img alt=\"...
2025-04-07T17:51:26
denoland/deno
018329a4d30681c953670cebe35a6e9dbf9bc5bb
c9c2531ff0209002be92c00d06f81262f2b2e641
fix: no cmd prefix in help output go links (#25459)
[ { "path": "cli/args/flags.rs", "patch": "@@ -1658,7 +1658,7 @@ Evaluate the given files, run all benches declared with 'Deno.bench()' and repor\n If you specify a directory instead of a file, the path is expanded to all contained files matching the glob <c>{*_,*.,}bench.{js,mjs,ts,mts,jsx,tsx}</>:\n <p(24...
2024-09-12T22:44:43
huggingface/transformers
3a826a45cab1ca84d3e2a5b9212b883bdb4ea74a
5e855095a222d5a2f9a8ef31f765aad2daa10bf1
Update Model card for GPT2 (#37101) * Update Model card for gpt2 * Update link for gpt2 space * fixes docs based on suggestions * Add transformers-cli and quantization example for GPT-2 * Remove resources and flash attention docs and fix typos
[ { "path": "docs/source/en/model_doc/gpt2.md", "patch": "@@ -14,197 +14,97 @@ rendered properly in your Markdown viewer.\n \n -->\n \n-# OpenAI GPT2\n-\n-<div class=\"flex flex-wrap space-x-1\">\n-<a href=\"https://huggingface.co/models?filter=gpt2\">\n-<img alt=\"Models\" src=\"https://img.shields.io/badge/...
2025-04-07T17:15:28
huggingface/transformers
5e855095a222d5a2f9a8ef31f765aad2daa10bf1
416b5a875d04a0268429f917645e014bb9b5d572
Update falcon mamba card (#37253) * feat: edit falcon mamba card * fix: edit statement on falconmamba arch * Update docs/source/en/model_doc/falcon_mamba.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/en/model_doc/falcon_mamba.md Co-authored-by: Steven Liu <5946235...
[ { "path": "docs/source/en/model_doc/falcon_mamba.md", "patch": "@@ -14,95 +14,100 @@ rendered properly in your Markdown viewer.\n \n -->\n \n-# FalconMamba\n-\n-<div class=\"flex flex-wrap space-x-1\">\n-<img alt=\"PyTorch\" src=\"https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor...
2025-04-07T17:12:44
denoland/deno
18b89d948dcb849c4dc577478794c3d5fb23b597
3f15e300625723df10c564c4e29ec276288a931b
fix(ext/node): Implement detached option in `child_process` (#25218) Fixes https://github.com/denoland/deno/issues/25193.
[ { "path": "ext/node/polyfills/internal/child_process.ts", "patch": "@@ -56,6 +56,7 @@ import { StringPrototypeSlice } from \"ext:deno_node/internal/primordials.mjs\";\n import { StreamBase } from \"ext:deno_node/internal_binding/stream_wrap.ts\";\n import { Pipe, socketType } from \"ext:deno_node/internal_b...
2024-09-12T19:24:58
denoland/deno
3f15e300625723df10c564c4e29ec276288a931b
0a4a8c730bcad32dcb7c8174e33601d777e54909
fix(ext/node): fix Decipheriv when autoPadding disabled (#25598) This change fixes Decipheriv behavior when autoPadding disabled and enabled. By this change, the example given in https://github.com/denoland/deno/issues/20924#issuecomment-2345931295 works in the same way as Node. closes #20924
[ { "path": "ext/node/polyfills/internal/crypto/cipher.ts", "patch": "@@ -306,6 +306,10 @@ class BlockModeCache {\n this.cache = this.cache.subarray(len);\n return out;\n }\n+\n+ set lastChunkIsNonZero(value: boolean) {\n+ this.#lastChunkIsNonZero = value;\n+ }\n }\n \n export class Decipheriv ...
2024-09-12T16:13:56
huggingface/transformers
f8a16805c5e96410f1fdd37733aa965e727f19d1
48e179857ce4574efec0210a7405eeb6f93b3974
updated model card for Mistral (#37156) * model card for Mistral * Update docs/source/en/model_doc/mistral.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Apply suggestions from code review Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/en...
[ { "path": "docs/source/en/model_doc/mistral.md", "patch": "@@ -14,74 +14,55 @@ rendered properly in your Markdown viewer.\n \n -->\n \n-# Mistral\n-\n-<div class=\"flex flex-wrap space-x-1\">\n-<img alt=\"PyTorch\" src=\"https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor=white\">\...
2025-04-07T17:05:36
denoland/deno
4983f763d480f73a7e0d55549afb7af87fa7f04b
c9065103b8ef0317fd8141f554fdc0a2f801e844
fix(ext/console): prevent duplicate error printing when the cause is assigned (#25327) This commit fixes the error format when the cause is assigned separately, ensuring that the cause is only printed once instead of twice. The fix addresses issue [#21651](https://github.com/denoland/deno/issues/21651).
[ { "path": "ext/console/01_console.js", "patch": "@@ -1287,6 +1287,7 @@ function getKeys(value, showHidden) {\n ArrayPrototypePushApply(keys, ArrayPrototypeFilter(symbols, filter));\n }\n }\n+ keys = ArrayPrototypeFilter(keys, (key) => key !== \"cause\");\n return keys;\n }\n ", "additions...
2024-09-12T12:02:32
huggingface/transformers
22065bd645e0f72774ddca9d40a87f48fe22df11
f789f960c87f7b9606be466fa04d4758d706e0fd
fix derived berts `_init_weights` (#37341) * fix derived berts * more * roformer
[ { "path": "src/transformers/models/albert/modeling_albert.py", "patch": "@@ -579,6 +579,8 @@ def _init_weights(self, module):\n elif isinstance(module, nn.LayerNorm):\n module.bias.data.zero_()\n module.weight.data.fill_(1.0)\n+ elif isinstance(module, AlbertMLMHead):\...
2025-04-07T16:25:07
denoland/deno
c9065103b8ef0317fd8141f554fdc0a2f801e844
11b5f2d3ea325a2e6792e4b75fddfcea8a593fb7
fix: add test ensuring als works across dynamic import (#25593) The fix is in https://github.com/denoland/deno_core/pull/888 Fixes: https://github.com/denoland/deno/issues/25275 Signed-off-by: snek <snek@deno.com>
[ { "path": "tests/unit_node/async_hooks_test.ts", "patch": "@@ -160,3 +160,15 @@ Deno.test(async function worksWithAsyncAPIs() {\n test();\n });\n });\n+\n+Deno.test(async function worksWithDynamicImports() {\n+ const store = new AsyncLocalStorage();\n+ // @ts-expect-error implicit any\n+ globalThis...
2024-09-12T05:42:26
huggingface/transformers
f789f960c87f7b9606be466fa04d4758d706e0fd
12bf24d6ae249add2e94559a63fc85785ef96326
Avoid build crashes when torch.version.xpu doesn't exist and fix Llama4 processor tests (#37346) * Avoid build crashes when torch.version.xpu doesn't exist * Trigger tests * Fix image token and skip inappropriate test * Remove ignore_errors=True * Add another skip
[ { "path": "src/transformers/testing_utils.py", "patch": "@@ -202,7 +202,7 @@\n \n IS_ROCM_SYSTEM = torch.version.hip is not None\n IS_CUDA_SYSTEM = torch.version.cuda is not None\n- IS_XPU_SYSTEM = torch.version.xpu is not None\n+ IS_XPU_SYSTEM = getattr(torch.version, \"xpu\", None) is not No...
2025-04-07T16:05:54
denoland/deno
46d5f38a0f839db34ecce67f3483469e82cb593d
3285801429eb7a40e022b571b7e3007c3e60250c
fix(ext/node): export request and response clases from `http2` module (#25592) Closes https://github.com/denoland/deno/issues/20612 Closes https://github.com/denoland/deno/issues/23326 This makes `qwik` work.
[ { "path": "ext/node/polyfills/http2.ts", "patch": "@@ -2295,7 +2295,7 @@ function onStreamTimeout(kind) {\n };\n }\n \n-class Http2ServerRequest extends Readable {\n+export class Http2ServerRequest extends Readable {\n readableEnded = false;\n \n constructor(stream, headers, options, rawHeaders) {\n@@...
2024-09-12T02:03:57
huggingface/transformers
12bf24d6ae249add2e94559a63fc85785ef96326
e7ad0770127c0e23e576681b10216f94c70ee1da
enable 2 llama UT cases on xpu (#37126) * enable tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits and tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits_bf16 on xpu Signed-off-by: YAO Matrix <matrix.yao@intel.com> * switch to use Expectations S...
[ { "path": "src/transformers/testing_utils.py", "patch": "@@ -202,9 +202,11 @@\n \n IS_ROCM_SYSTEM = torch.version.hip is not None\n IS_CUDA_SYSTEM = torch.version.cuda is not None\n+ IS_XPU_SYSTEM = torch.version.xpu is not None\n else:\n IS_ROCM_SYSTEM = False\n IS_CUDA_SYSTEM = False\n+...
2025-04-07T14:02:14
denoland/deno
025e4bdaaca3ce029602ba3fef63865f6573e5e4
f794781ffba941085343b95dd5a9c1ad74a322a2
fix(ext/webgpu): sync category comment (#25580) Signed-off-by: Victor Turansky <victor.turansky@gmail.com>
[ { "path": "cli/tsc/dts/lib.deno_webgpu.d.ts", "patch": "@@ -600,16 +600,10 @@ declare interface GPUPipelineErrorInit {\n reason: GPUPipelineErrorReason;\n }\n \n-/**\n- * @category GPU\n- * @\n- */\n+/** @category GPU */\n declare type GPUPipelineErrorReason = \"validation\" | \"internal\";\n \n-/**\n- * ...
2024-09-12T00:59:05
huggingface/transformers
99f9f1042f59f60de9a8f0538c1117e4eca38ef9
0fb8d49e8869fe95c7580ae0b12cca41e009fe3f
Fix torchao usage (#37034) * fix load path Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix path Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * Fix torchao usage Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix tests Signed-off-by: jiqing-feng <jiqing.feng@intel.com> * fix format Signed-...
[ { "path": "src/transformers/utils/quantization_config.py", "patch": "@@ -20,7 +20,7 @@\n import importlib.metadata\n import json\n import os\n-from dataclasses import dataclass\n+from dataclasses import dataclass, is_dataclass\n from enum import Enum\n from inspect import Parameter, signature\n from typing ...
2025-04-07T12:50:48
huggingface/transformers
08f36771b33d246986d9338a729fc4ef258b999d
9db31ea58579cf441bc0cf978ecf917a289fdc39
Fix `init empty weights` without accelerate (#37337) * add the integration * Update accelerate.py * Update accelerate.py * add find_tied_params as well * Update accelerate.py * add where copied from * simplify * add error
[ { "path": "src/transformers/integrations/accelerate.py", "patch": "@@ -0,0 +1,196 @@\n+# Copyright 2025 The HuggingFace Team. All rights reserved.\n+#\n+# Licensed under the Apache License, Version 2.0 (the \"License\");\n+# you may not use this file except in compliance with the License.\n+# You may obtain...
2025-04-07T09:37:29
denoland/deno
c64aa50c0e4219ef623a6ec3b939ac10a4568563
aae3a6bcb41fdb305d61a95f176be0db1513c56e
feat(upgrade): better error message on failure (#25503) Co-authored-by: crowlkats <crowlkats@toaxl.com>
[ { "path": "cli/args/flags.rs", "patch": "@@ -2986,12 +2986,8 @@ The declaration file could be saved and used for typing information.\",\n )\n }\n \n-fn upgrade_subcommand() -> Command {\n- command(\n- \"upgrade\",\n- cstr!(\"Upgrade deno executable to the given version.\n-\n-<g>Latest</>\n+pub stat...
2024-09-11T15:19:45
huggingface/transformers
9db31ea58579cf441bc0cf978ecf917a289fdc39
debfe904c94d1104e94c71d2276a0d522f0f5081
Fix deepspeed with quantization (#37324) * Update modeling_utils.py * Update modeling_utils.py
[ { "path": "src/transformers/modeling_utils.py", "patch": "@@ -3719,19 +3719,14 @@ def float(self, *args):\n \n @classmethod\n def get_init_context(cls, is_quantized: bool, _is_ds_init_called: bool):\n- # With deepspeed, we cannot initialize the model on meta device\n if is_deepspeed_z...
2025-04-07T09:36:44
denoland/deno
aae3a6bcb41fdb305d61a95f176be0db1513c56e
200145a09a51e96298f1ecd5ce78f75c27c7880a
fix(ext/node): fix `Cipheriv#update(string, undefined)` (#25571)
[ { "path": "ext/node/polyfills/internal/crypto/cipher.ts", "patch": "@@ -242,7 +242,7 @@ export class Cipheriv extends Transform implements Cipher {\n ): Buffer | string {\n // TODO(kt3k): throw ERR_INVALID_ARG_TYPE if data is not string, Buffer, or ArrayBufferView\n let buf = data;\n- if (typeo...
2024-09-11T10:24:17
huggingface/transformers
debfe904c94d1104e94c71d2276a0d522f0f5081
54538ebee3005951e62c16d2e9235cb5dfa19f83
fix llama4 training (#37319)
[ { "path": "src/transformers/models/llama4/modeling_llama4.py", "patch": "@@ -356,7 +356,7 @@ def forward(\n attn_scales = (\n torch.log(torch.floor((cache_position.float() + 1.0) / self.floor_scale) + 1.0) * self.attn_scale + 1.0\n )\n- attn_scales = attn_s...
2025-04-07T07:24:44
denoland/deno
200145a09a51e96298f1ecd5ce78f75c27c7880a
ad30703e8e6bcf0fb87b5bc4ad1b49d040e54c77
fix(ext/node): avoid showing `UNKNOWN` error from TCP handle (#25550)
[ { "path": "ext/node/polyfills/internal_binding/stream_wrap.ts", "patch": "@@ -38,6 +38,7 @@ import { TextEncoder } from \"ext:deno_web/08_text_encoding.js\";\n import { Buffer } from \"node:buffer\";\n import { notImplemented } from \"ext:deno_node/_utils.ts\";\n import { HandleWrap } from \"ext:deno_node/i...
2024-09-11T10:19:02
huggingface/transformers
54538ebee3005951e62c16d2e9235cb5dfa19f83
d1b92369ca193da49f9f7ecd01b08ece45c2c9aa
fix flex attn when optional args aren't passed (#37327)
[ { "path": "src/transformers/integrations/flex_attention.py", "patch": "@@ -100,6 +100,11 @@ def make_flex_block_causal_mask(\n Returns:\n BlockMask\n \"\"\"\n+ batch_size, total_seq_len = attention_mask_2d.shape\n+ if not key_length:\n+ key_length = total_seq_len\n+ if not qu...
2025-04-07T07:12:21
denoland/deno
ad30703e8e6bcf0fb87b5bc4ad1b49d040e54c77
8bdd364dd568f93097ecee41e66c74d16d75c015
fix(lsp): encode url parts before parsing as uri (#25509)
[ { "path": "cli/lsp/diagnostics.rs", "patch": "@@ -12,6 +12,7 @@ use super::language_server::StateSnapshot;\n use super::performance::Performance;\n use super::tsc;\n use super::tsc::TsServer;\n+use super::urls::uri_parse_unencoded;\n use super::urls::url_to_uri;\n use super::urls::LspUrlMap;\n \n@@ -53,11 +...
2024-09-11T10:11:39
denoland/deno
8bdd364dd568f93097ecee41e66c74d16d75c015
87bc47b3bf9afe26e33450757b13c3b14b8703d8
fix(ext/node): add `FileHandle#writeFile` (#25555) This PR adds `writeFile` methods of `FileHandle` class https://nodejs.org/api/fs.html#filehandlewritefiledata-options
[ { "path": "ext/node/polyfills/_fs/_fs_writeFile.ts", "patch": "@@ -23,14 +23,15 @@ import {\n validateStringAfterArrayBufferView,\n } from \"ext:deno_node/internal/fs/utils.mjs\";\n import { promisify } from \"ext:deno_node/internal/util.mjs\";\n+import { FileHandle } from \"ext:deno_node/internal/fs/hand...
2024-09-11T09:08:40
huggingface/transformers
25b7f272347a93d6fb73cad126f6f6dc88e8ce89
aa40fda346b497435923ee1ac2120900653f5ab5
Add llama4 (#37307) * remove one of the last deps * update fast image processor after refactor * styling * more quality of life improvements * nit * update * cleanups * some cleanups * vllm updates * update fake image token * [convert] Fix typo * [convert] Strip extraneous bytes from shards * [convert] Min...
[ { "path": "docs/source/en/_toctree.yml", "patch": "@@ -507,6 +507,8 @@\n title: Llama2\n - local: model_doc/llama3\n title: Llama3\n+ - local: model_doc/llama4\n+ title: Llama4\n - local: model_doc/longformer\n title: Longformer\n - local: model_doc/long...
2025-04-05T20:02:22
denoland/deno
87bc47b3bf9afe26e33450757b13c3b14b8703d8
ef2d98fe11ffe467a31d2e30e3ae9738147b74e9
fix(ext/node): use primordials in `ext/node/polyfills/console.ts` (#25572) Contributing toward #24236 - Swapped `Object.assign` for `ObjectAssign` primordial. - Removed referencing TODO comment. Please disregard if no longer desired.
[ { "path": "ext/node/polyfills/console.ts", "patch": "@@ -1,15 +1,15 @@\n // Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.\n \n-// TODO(petamoriken): enable prefer-primordials for node polyfills\n-// deno-lint-ignore-file prefer-primordials\n-\n+import { primordials } from \"ext:cor...
2024-09-11T07:55:58
huggingface/transformers
e94571580ba9d0c71feaf489520383c83e167d40
84aa13dd85ce5ec2023561ca304c5b41343dd347
Fix deepspeed loading (part 2) (#37306) * fix * Update modeling_utils.py * Update modeling_utils.py * oups remove print
[ { "path": "src/transformers/modeling_utils.py", "patch": "@@ -3723,18 +3723,22 @@ def float(self, *args):\n \n @classmethod\n def get_init_context(cls, is_quantized: bool, _is_ds_init_called: bool):\n- if is_deepspeed_zero3_enabled() and not is_quantized and not _is_ds_init_called:\n- ...
2025-04-05T18:41:42
denoland/deno
ef2d98fe11ffe467a31d2e30e3ae9738147b74e9
1521adf5ed640832755e362abc64b32afd7dcc7d
fix(ext/node): validate input lengths in `Cipheriv` and `Decipheriv` (#25570) addresses the first part of #25279
[ { "path": "ext/node/ops/crypto/cipher.rs", "patch": "@@ -4,6 +4,7 @@ use aes::cipher::block_padding::Pkcs7;\n use aes::cipher::BlockDecryptMut;\n use aes::cipher::BlockEncryptMut;\n use aes::cipher::KeyIvInit;\n+use deno_core::error::range_error;\n use deno_core::error::type_error;\n use deno_core::error::A...
2024-09-11T04:27:07
huggingface/transformers
84aa13dd85ce5ec2023561ca304c5b41343dd347
0ef339ff1b63bb03a388c79bfbebec9085e10564
Fix deepspeed loading (#37281) * Update modeling_utils.py * Update modeling_utils.py * fix and remove all imports * Update modeling_utils.py * Update modeling_utils.py * style * Update modeling_utils.py
[ { "path": "src/transformers/modeling_utils.py", "patch": "@@ -57,7 +57,7 @@\n from .dynamic_module_utils import custom_object_save\n from .generation import CompileConfig, GenerationConfig, GenerationMixin\n from .integrations import PeftAdapterMixin, deepspeed_config, is_deepspeed_zero3_enabled\n-from .int...
2025-04-05T15:05:45
denoland/deno
1521adf5ed640832755e362abc64b32afd7dcc7d
acd62786c24c38d8986475f0660d5205b55a9ad3
fix(ext/node): Add missing node:fs and node:constants exports (#25568) Just a bunch of random fs constants
[ { "path": "ext/node/polyfills/_fs/_fs_constants.ts", "patch": "@@ -12,6 +12,28 @@ export const {\n S_IXUSR,\n S_IRGRP,\n S_IWGRP,\n+ S_IFBLK,\n+ S_IFCHR,\n+ S_IFDIR,\n+ S_IFIFO,\n+ S_IFLNK,\n+ S_IFMT,\n+ S_IFREG,\n+ S_IFSOCK,\n+ S_IRWXG,\n+ S_IRWXO,\n+ S_IRWXU,\n+ UV_DIRENT_BLOCK,\n+ UV_D...
2024-09-11T00:37:23
huggingface/transformers
46d73910d5922c154967309c26fe8460a505eb3f
579135a2f64be5b0727ceba696e5642ba3612779
Updated T5 model card with standardized format (#37261) * Updated T5 model card with standardized format * Updated T5 model card with standardized format, fixed typo * Update docs/source/en/model_doc/t5.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/en/model_doc/t5....
[ { "path": "docs/source/en/model_doc/t5.md", "patch": "@@ -14,349 +14,104 @@ rendered properly in your Markdown viewer.\n \n -->\n \n-# T5\n-\n-<div class=\"flex flex-wrap space-x-1\">\n-<img alt=\"PyTorch\" src=\"https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor=white\">\n-<img a...
2025-04-04T22:23:09
denoland/deno
acd62786c24c38d8986475f0660d5205b55a9ad3
f959297dcd382b25bdedd0ff6aa27e8fb40e7ecd
fix(ext/node): Add missing `node:path` exports (#25567) Apparently `path/posix` and `path/win32` have circular exports. I do not know why. Additionally there's a deprecated function `_makeLong` which is just `toNamespacedPath`
[ { "path": "ext/node/polyfills/path/_posix.ts", "patch": "@@ -478,6 +478,9 @@ export function parse(path: string): ParsedPath {\n \n return ret;\n }\n+\n+export const _makeLong = toNamespacedPath;\n+\n export default {\n basename,\n delimiter,\n@@ -492,4 +495,5 @@ export default {\n resolve,\n sep,...
2024-09-11T00:00:08
huggingface/transformers
579135a2f64be5b0727ceba696e5642ba3612779
8cd57eb73107de0353d1fe94cfe99484b738172a
Updated model card for distilbert (#37157) * Updated model card for distilbert * Updated the distilbert model card * Updated model card for distilbert * Updated the distilbert model card * Addressed code review comments * Addressed review comments * fix pipeline --------- Co-authored-by: Steven Liu <59462357+s...
[ { "path": "docs/source/en/model_doc/distilbert.md", "patch": "@@ -14,199 +14,91 @@ rendered properly in your Markdown viewer.\n \n -->\n \n-# DistilBERT\n-\n-<div class=\"flex flex-wrap space-x-1\">\n-<img alt=\"PyTorch\" src=\"https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor=wh...
2025-04-04T22:22:46