Defined in 28 files
Last patched
__init__ in huggingface/transformers was changed by 6 published CVE fixes between 2023 and 2026, in 28 distinct patches that added or removed 128 lines in total.
The name occurs in 28 files: src/transformers/models/deprecated/transfo_xl/tokenization_transfo_xl.py, src/transformers/models/pvt_v2/modeling_pvt_v2.py, src/transformers/models/timesformer/modeling_timesformer.py, src/transformers/models/gpt_neox_japanese/tokenization_gpt_neox_japanese.py, src/transformers/models/mgp_str/modeling_mgp_str.py, src/transformers/models/cvt/modeling_cvt.py, src/transformers/models/lightglue/modeling_lightglue.py, src/transformers/models/glpn/modeling_glpn.py, src/transformers/models/swin/modeling_swin.py, src/transformers/models/segformer/modeling_segformer.py, src/transformers/models/lightglue/modular_lightglue.py, src/transformers/models/donut/modeling_donut_swin.py, src/transformers/models/data2vec/modeling_data2vec_vision.py, src/transformers/models/seggpt/modeling_seggpt.py, src/transformers/models/swinv2/modeling_swinv2.py, src/transformers/models/convnext/modeling_convnext.py, src/transformers/models/focalnet/modeling_focalnet.py, src/transformers/models/poolformer/modeling_poolformer.py, src/transformers/models/swin2sr/modeling_swin2sr.py, src/transformers/models/convnextv2/modeling_convnextv2.py, src/transformers/models/hiera/modeling_hiera.py, src/transformers/models/vitdet/modeling_vitdet.py, src/transformers/models/deprecated/gptsan_japanese/tokenization_gptsan_japanese.py, src/transformers/models/maskformer/modeling_maskformer_swin.py, src/transformers/models/pvt/modeling_pvt.py, src/transformers/models/clap/modeling_clap.py, src/transformers/configuration_utils.py, src/transformers/models/beit/modeling_beit.py.
huggingface/transformers has 21 CVEs with a located fix in total, so 29% of them touched this function.
Every published CVE whose fix changed __init__, newest first, with the lines each fix touched.
CVE-2026-5241 · src/transformers/models/lightglue/modeling_lightglue.py · line 505 · +1 −3
CVE-2026-5241 · src/transformers/models/lightglue/modular_lightglue.py · line 512 · +1 −3
CVE-2026-4372 · src/transformers/configuration_utils.py · lines 266–272 · +7 −5
CVE-2025-3777 · src/transformers/models/beit/modeling_beit.py · line 666 · +1 −1
CVE-2025-3777 · src/transformers/models/clap/modeling_clap.py · line 832 · +1 −1
CVE-2025-3777 · src/transformers/models/convnext/modeling_convnext.py · lines 228–229 · +2 −1
CVE-2025-3777 · src/transformers/models/convnextv2/modeling_convnextv2.py · lines 248–249 · +2 −1
CVE-2025-3777 · src/transformers/models/cvt/modeling_cvt.py · lines 452–454 · +3 −1
CVE-2025-3777 · src/transformers/models/data2vec/modeling_data2vec_vision.py · line 679 · +1 −1
CVE-2025-3777 · src/transformers/models/donut/modeling_donut_swin.py · line 793 · +1 −1
CVE-2025-3777 · src/transformers/models/focalnet/modeling_focalnet.py · line 489 · +1 −1
CVE-2025-3777 · src/transformers/models/glpn/modeling_glpn.py · line 334 · +1 −1
CVE-2025-3777 · src/transformers/models/hiera/modeling_hiera.py · lines 642–644 · +2 −2
CVE-2025-3777 · src/transformers/models/maskformer/modeling_maskformer_swin.py · line 695 · +1 −1
CVE-2025-3777 · src/transformers/models/mgp_str/modeling_mgp_str.py · line 249 · +1 −1
CVE-2025-3777 · src/transformers/models/poolformer/modeling_poolformer.py · line 197 · +1 −1
CVE-2025-3777 · src/transformers/models/pvt/modeling_pvt.py · line 372 · +1 −1
CVE-2025-3777 · src/transformers/models/pvt_v2/modeling_pvt_v2.py · line 326 · +1 −1
CVE-2025-3777 · src/transformers/models/segformer/modeling_segformer.py · lines 359–361 · +3 −1
CVE-2025-3777 · src/transformers/models/seggpt/modeling_seggpt.py · line 463 · +1 −1
CVE-2025-3777 · src/transformers/models/swin2sr/modeling_swin2sr.py · line 685 · +1 −1
CVE-2025-3777 · src/transformers/models/swin/modeling_swin.py · line 826 · +1 −1
CVE-2025-3777 · src/transformers/models/swinv2/modeling_swinv2.py · line 880 · +1 −1
CVE-2025-3777 · src/transformers/models/timesformer/modeling_timesformer.py · line 298 · +1 −1
CVE-2025-3777 · src/transformers/models/vitdet/modeling_vitdet.py · line 538 · +1 −1
CVE-2025-1194 · src/transformers/models/deprecated/gptsan_japanese/tokenization_gptsan_japanese.py · lines 411–427 · +17 −3
CVE-2025-1194 · src/transformers/models/gpt_neox_japanese/tokenization_gpt_neox_japanese.py · lines 234–250 · +17 −3
CVE-2023-7018 · src/transformers/models/deprecated/transfo_xl/tokenization_transfo_xl.py · lines 216–223 · +8 −0
CVE-2023-6730 · src/transformers/models/deprecated/transfo_xl/tokenization_transfo_xl.py · lines 216–223 · +8 −0
Get an alert when the next CVE fix lands in huggingface/transformers, with the patched functions.
Monitor transformers CVEsFunctions in the same files that CVE fixes also changed.
Functions in huggingface/transformers changed by the most CVE fixes.
Common questions about CVE fixes to __init__.
Fix locations are computed by RadicalNotion.AI from the upstream patch diff of each CVE. CVE data is sourced from the CVE Program.
Get an alert when a new CVE lands in this repository, with the patched functions and AI-written remediation guidance.
