44 repos · last 2 days of GitHub activity · 659 monkey patches tracked
| action | component | what | threads |
|---|
| thread | comments | title |
|---|
Something changed on their side or ours: check whether a transformers change forced it.
None.
Patched by several repos = missing extension point (kernel hub / attention interface / hook).
| symbol | #repos | repos |
|---|---|---|
*Attention.forward | 8 | FlagEmbedding, TensorRT-LLM, axolotl, LlamaFactory, ms-swift, unsloth, unsloth-zoo, verl |
*ForCausalLM.forward | 6 | Automodel, axolotl, Liger-Kernel, ms-swift, unsloth, unsloth-zoo |
*Model.forward | 6 | lmms-eval, axolotl, ms-swift, unsloth, unsloth-zoo, verl |
*MLP.forward | 5 | TensorRT-LLM, axolotl, DeepSpeed, unsloth, verl |
*DecoderLayer.forward | 4 | axolotl, LlamaFactory, unsloth, verl |
*FlashAttention2 | 4 | GPTQModel, ms-swift, sglang, unsloth-zoo |
*Experts | 3 | axolotl, Liger-Kernel, unsloth-zoo |
*Model._create_and_replace | 3 | axolotl, ms-swift, unsloth-zoo |
*RMSNorm | 3 | axolotl, Liger-Kernel, unsloth |
PreTrainedModel.from_pretrained | 3 | ColossalAI, ms-swift, unsloth |
apply_rotary_pos_emb | 3 | axolotl, Liger-Kernel, verl |
create_causal_mask | 3 | GPTQModel, ms-swift, unsloth-zoo |
*Experts.forward | 2 | unsloth-zoo, verl |
*FlashAttention2.forward | 2 | LlamaFactory, unsloth |
*MLP | 2 | axolotl, Liger-Kernel |
*Model._postprocess | 2 | ROLL, verl |
*Model.load_adapter | 2 | ms-swift, unsloth-zoo |
*Model.save_pretrained | 2 | unsloth, unsloth-zoo |
*RMSNorm.forward | 2 | Automodel, verl |
*SdpaAttention.forward | 2 | LlamaFactory, unsloth |
*SparseMoeBlock | 2 | LlamaFactory, verl |
CrossEntropyLoss | 2 | axolotl, Liger-Kernel |
PreTrainedModel.enable_input_require_grads | 2 | ms-swift, unsloth |
PreTrainedModel.post_init | 2 | sglang, unsloth |
_init_weights | 2 | GPTQModel, ColossalAI |
checkpoint | 2 | ms-swift, unsloth-zoo |
dynamic_module_utils.check_imports | 2 | LlamaFactory, ktransformers |
get_model_conversion_mapping | 2 | unsloth, unsloth-zoo |
nn.CrossEntropyLoss | 2 | axolotl, Liger-Kernel |
nn.LayerNorm | 2 | axolotl, Liger-Kernel |
*Attention | 1 | GPTQModel |
*Attention.ATTN_DONE | 1 | ColossalAI |
*Attention.CU_SEQLENS | 1 | ColossalAI |
*Attention.DEFAULT_BLOCK_SIZES | 1 | diffusers |
*Attention.HALF_INDICES | 1 | ColossalAI |
*Attention.INNER_RING_GROUP | 1 | ColossalAI |
*Attention.INTER_RING_GROUP | 1 | ColossalAI |
*Attention.SP_GROUP | 1 | ColossalAI |
*Attention.SP_STREAM | 1 | ColossalAI |
*Attention.TOTAL_SEQLEN | 1 | ColossalAI |
| where | patch |
|---|---|
| EvolvingLMMs-Lab/lmms-eval/lmms_eval/models/simple/cambrians_vsc.py | CambrianQwenModel.forward = cambrian_qwen2_forward |
| EvolvingLMMs-Lab/lmms-eval/lmms_eval/models/simple/cambrians_vsr.py | CambrianQwenModel.forward = cambrian_qwen2_forward |
| FlagOpen/FlagEmbedding/research/llm_embedder/src/utils/llama_patch.py | transformers.models.llama.modeling_llama.LlamaAttention.forward = forward |
| FlagOpen/FlagEmbedding/research/llm_embedder/src/utils/llama_patch.py | transformers.models.llama.modeling_llama.LlamaModel._prepare_decoder_attention_mask = ( |
| ModelCloud/GPTQModel/gptqmodel/models/base.py | modeling_utils.check_support_param_buffer_assignment = check_support_param_buffer_assignment |
| ModelCloud/GPTQModel/gptqmodel/models/definitions/intern_s2_preview.py | modeling_module.create_causal_mask = create_causal_mask_compat |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | PreTrainedModel.__getattr__ = __getattr__ |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | PreTrainedModel._gptqmodel_legacy_tied_weights_patch = True |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | PreTrainedModel._gptqmodel_missing_all_tied_weights_patch = True |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | PreTrainedModel.get_expanded_tied_weights_keys = get_expanded_tied_weights_keys |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | PreTrainedModel.get_head_mask = get_head_mask |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | PreTrainedModel.is_parallelizable = False |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_auto.AutoModelForCausalLM.register = classmethod(patched_register) |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_llama.LlamaAttention = DeepSeekVLV2CompatLlamaAttention |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_llama.LlamaFlashAttention2 = DeepSeekVLV2CompatLlamaAttention |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_llama.LlamaFlashAttention2 = original_llama_attention |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_utils._init_weights = False |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_utils._init_weights = original_init_flag |
| ModelCloud/GPTQModel/gptqmodel/utils/hf.py | modeling_utils.is_flash_attn_2_available = is_flash_attn_2_available |
| NVIDIA-NeMo/Automodel/nemo_automodel/_transformers/auto_model.py | PreTrainedModel._finalize_model_loading = staticmethod(_compat_finalize) |
| NVIDIA-NeMo/Automodel/nemo_automodel/_transformers/model_init.py | PreTrainedModel.get_init_context = classmethod(_patched_get_init_context) |
| NVIDIA-NeMo/Automodel/nemo_automodel/_transformers/utils.py | Qwen3ForCausalLM.forward = _patched_forward # type: ignore[method-assign] |
| NVIDIA-NeMo/Automodel/nemo_automodel/components/models/beam/model.py | BeamForCausalLM.parallelizer = PARALLELIZER |
| NVIDIA-NeMo/Automodel/nemo_automodel/components/models/common/utils.py | TERMSNorm.forward = _patched_rmsnorm_forward |
| NVIDIA-NeMo/Automodel/nemo_automodel/components/models/deepseek_v4/model.py | DeepseekV4ForCausalLM.parallelizer = PARALLELIZER |
| NVIDIA-NeMo/Automodel/nemo_automodel/components/models/deepseek_v41/model.py | DeepseekV41ForCausalLM.parallelizer = PARALLELIZER |
| NVIDIA-NeMo/Automodel/nemo_automodel/components/models/nemotron_v3/model.py | NemotronHForCausalLM.parallelizer = PARALLELIZER |
| NVIDIA-NeMo/Automodel/nemo_automodel/components/models/qwen3_5/model.py | Qwen3_5ForCausalLM.parallelizer = PARALLELIZER |
| NVIDIA-NeMo/Automodel/nemo_automodel/shared/transformers_patches.py | modeling_t5.T5LayerNorm = _NativeT5LayerNorm |
| NVIDIA/TensorRT-LLM/tensorrt_llm/tools/layer_wise_benchmarks/mark_utils.py | GatedMLP.forward = nvtx.annotate("GatedMLP")(GatedMLP.forward) |
| NVIDIA/TensorRT-LLM/tensorrt_llm/tools/layer_wise_benchmarks/mark_utils.py | KimiKDALinearAttention.forward = nvtx.annotate("KimiKDALinearAttention")( |
| OpenRLHF/OpenRLHF/openrlhf/utils/deepspeed/deepspeed.py | transformers.modeling_flash_attention_utils.deterministic_g = True |
| PrimeIntellect-ai/prime-rl/src/prime_rl/inference/patches.py | Glm5NextForCausalLM.process_weights_after_loading = process_weights_after_loading |
| PrimeIntellect-ai/prime-rl/src/prime_rl/inference/patches.py | LoRAModel.from_local_checkpoint = classmethod(_patched_from_local_checkpoint) |
| PrimeIntellect-ai/prime-rl/src/prime_rl/inference/patches.py | MiniMaxM2ForCausalLM.hf_to_vllm_mapper = WeightsMapper( |
| PrimeIntellect-ai/prime-rl/src/prime_rl/inference/vllm/qwen38_weight_loading.py | Qwen4ExpNGramEmbedding.load_weights = load_weights |
| PrimeIntellect-ai/prime-rl/src/prime_rl/trainer/models/gpt_oss/attention.py | GptOssAttention.compute_attention = ulysses_attention |
| PrimeIntellect-ai/prime-rl/src/prime_rl/trainer/models/layers/ulysses_attn.py | AfmoeFlashAttention._compute_attention = _ulysses_compute_attention |
| PrimeIntellect-ai/prime-rl/src/prime_rl/trainer/models/layers/ulysses_attn.py | FlashAttention._compute_attention = _ulysses_compute_attention |
| alibaba/ROLL/roll/models/model_providers.py | AutoModelForCausalLMWithValueHead.post_init = no_set_device_hook_post_init |
| alibaba/ROLL/roll/third_party/megatron/mtp_patcher.py | GPTModel._original_postprocess = original_gpt_postprocess |
| alibaba/ROLL/roll/third_party/megatron/mtp_patcher.py | GPTModel._postprocess = patched_gpt_postprocess |
| alibaba/ROLL/roll/third_party/vllm/gdn_patcher.py | GDNAttentionMetadataBuilder._original_build = original_build |
| alibaba/ROLL/roll/third_party/vllm/gdn_patcher.py | GDNAttentionMetadataBuilder.build = patched_build |
| alibaba/ROLL/roll/utils/context_parallel/monkey_patch.py | Qwen2Model._update_causal_mask = _update_causal_mask |
| alibaba/ROLL/roll/utils/context_parallel/monkey_patch.py | Qwen2Model._update_causal_mask = old_update_causal_mask |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/expert_parallel/experts_fn.py | PreTrainedModel.get_correct_experts_implementation = patched # type: ignore[assignment] |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/kernels/libs/scattermoe_lora/experts.py | PreTrainedModel.get_correct_experts_implementation = _patched_get_correct |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/kernels/libs/scattermoe_lora/nvfp4_moe_loading.py | PreTrainedModel._initialize_missing_keys = patched |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/glm4_moe_lite.py | modeling_glm4_moe_lite.Glm4MoeLiteMLP.forward = LigerSwiGLUMLP.forward |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/glm4_moe_lite.py | modeling_glm4_moe_lite.Glm4MoeLiteRMSNorm = LigerRMSNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/glm4_moe_lite.py | modeling_glm4_moe_lite.apply_rotary_pos_emb = liger_rotary_pos_emb |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/glm4_moe_lite.py | modeling_glm4_moe_lite.apply_rotary_pos_emb_interleave = ( |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5.py | modeling_qwen3_5.Qwen3_5ForCausalLM.forward = lce_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5.py | modeling_qwen3_5.Qwen3_5MLP = _liger_swiglu_mlp_wrapper |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5.py | modeling_qwen3_5.Qwen3_5RMSNorm = LigerRMSNormForQwen3_5 |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5.py | modeling_qwen3_5.Qwen3_5RMSNormGated = FusedRMSNormGated |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5.py | modeling_qwen3_5.nn.LayerNorm = LigerLayerNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5_moe.py | modeling_mod.Qwen3_5MoeForCausalLM.forward = lce_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5_moe.py | modeling_mod.Qwen3_5MoeMLP = _liger_swiglu_mlp_wrapper |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5_moe.py | modeling_mod.Qwen3_5MoeRMSNorm = LigerRMSNormForQwen3_5Moe |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5_moe.py | modeling_mod.Qwen3_5MoeRMSNormGated = FusedRMSNormGated |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/models/qwen3_5_moe.py | modeling_mod.nn.LayerNorm = LigerLayerNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_cohere_compass.CohereCompassMLP = LigerSwiGLUMLP |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_cohere_compass.nn.LayerNorm = LigerLayerNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_gemma4_unified.Gemma4UnifiedRMSNorm = ( |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_gemma4_unified.Gemma4UnifiedTextMLP = _LigerGemma4UnifiedMLP |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_gemma4_unified.nn.CrossEntropyLoss = LigerCrossEntropyLoss |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_gemma4_unified.nn.LayerNorm = LigerLayerNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_jamba.JambaForCausalLM.forward = jamba_lce_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_jamba.JambaMLP = LigerSwiGLUMLP |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_jamba.JambaRMSNorm = LigerRMSNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_jamba.apply_rotary_pos_emb = liger_rotary_pos_emb |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_jamba.nn.LayerNorm = LigerLayerNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_ministral3.Ministral3MLP = LigerSwiGLUMLP |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_ministral3.Ministral3RMSNorm = LigerRMSNorm |
| axolotl-ai-cloud/axolotl/src/axolotl/integrations/liger/plugin.py | modeling_ministral3.apply_rotary_pos_emb = liger_rotary_pos_emb |
| axolotl-ai-cloud/axolotl/src/axolotl/loaders/adapter.py | LoraModel._axolotl_clippable_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/loaders/adapter.py | LoraModel._create_and_replace = _patched |
| axolotl-ai-cloud/axolotl/src/axolotl/loaders/model.py | transformers.integrations.deepspeed.is_deepspeed_zero3_enabled = lambda: ( |
| axolotl-ai-cloud/axolotl/src/axolotl/loaders/model.py | transformers.modeling_utils.is_deepspeed_zero3_enabled = lambda: True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/accelerate/fsdp2.py | PreTrainedModel._adjust_tied_keys_with_tied_pointers = ( |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/accelerate/fsdp2.py | PreTrainedModel._initialize_missing_keys = _patched_initialize_missing_keys |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/accelerate/fsdp2.py | PreTrainedModel._move_missing_keys_from_meta_to_device = _patched_move_missing_keys |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/attention/flex_attn.py | transformers.integrations.flex_attention.WrappedFlexAttention = WrappedFlexAttention |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/btlm_attn_hijack_flash.py | modeling_btlm.BTLMAttention._attn = flashattn_attn |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/llama_attn_hijack_flash.py | transformers.loss.loss_utils.fixed_cross_entropy = fa2_fixed_cross_entropy |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/llama_attn_hijack_xformers.py | transformers.models.llama.modeling_llama.LlamaAttention.forward = xformers_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/loss/chunked.py | transformers.loss.loss_utils.ForCausalLMLoss = for_causal_lm_chunked_loss |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/mistral_attn_hijack_flash.py | transformers.models.mistral.modeling_mistral.CrossEntropyLoss = partial( |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/mixtral/__init__.py | MixtralBlockSparseTop2MLP.forward = mlp_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/falcon_h1/modeling.py | FalconH1DecoderLayer.forward = patched_decoder_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4/fused_attn.py | Gemma4TextAttention.forward = _make_fused_forward(original_forward) |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4/fused_attn.py | Gemma4TextDecoderLayer.__call__ = patched_call |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4/fused_attn.py | Gemma4TextDecoderLayer._axolotl_shared_kv_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4_unified/fused_attn.py | Gemma4UnifiedTextAttention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4_unified/fused_attn.py | Gemma4UnifiedTextAttention.forward = _make_fused_forward(original_forward) |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4_unified/fused_attn.py | Gemma4UnifiedTextDecoderLayer.__call__ = patched_call |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/gemma4_unified/fused_attn.py | Gemma4UnifiedTextDecoderLayer._axolotl_shared_kv_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/glm4_moe_lite/fused_attn.py | Glm4MoeLiteAttention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/glm4_moe_lite/fused_attn.py | Glm4MoeLiteAttention.forward = _make_fused_forward(Glm4MoeLiteAttention.forward) |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/llama4/modeling.py | modeling_llama4.Llama4TextExperts = Llama4TextExperts |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/llama4/modeling.py | modeling_llama4.Llama4TextExperts = old_lamma_4_text_experts |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/pixtral/modeling_flash_attention_utils.py | modeling_flash_attention_utils._is_packed_sequence = fixed_is_packed_sequence |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/pixtral/modeling_flash_attention_utils.py | modeling_flash_attention_utils._is_packed_sequence = old_fn |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3/fused_attn.py | Qwen3Attention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3/fused_attn.py | Qwen3Attention.forward = _make_fused_forward() |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_5/fused_attn.py | Qwen3_5Attention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_5/fused_attn.py | Qwen3_5Attention.forward = _make_fused_forward() |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_5_moe/fused_attn.py | Qwen3_5MoeAttention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_5_moe/fused_attn.py | Qwen3_5MoeAttention.forward = _make_fused_forward() |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_moe/fused_attn.py | Qwen3MoeAttention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_moe/fused_attn.py | Qwen3MoeAttention.forward = _make_fused_forward() |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_next/modeling.py | Qwen3NextDecoderLayer.forward = original_decoder_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_next/modeling.py | Qwen3NextDecoderLayer.forward = patched_decoder_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_vl/fused_attn.py | Qwen3VLTextAttention._axolotl_fused_attn_patched = True |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/models/qwen3_vl/fused_attn.py | Qwen3VLTextAttention.forward = _make_fused_forward() |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/moe_quant.py | transformers.core_model_loading.set_param_for_module = _patched_set_param_for_module |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/moe_quant.py | transformers.modeling_utils.caching_allocator_warmup = _noop_warmup |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/multipack.py | modeling_arch._get_unpad_data = get_unpad_data |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/peft/nf4.py | LoraModel.merge_and_unload = merge_and_unload |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/quantized_init.py | PreTrainedModel._initialize_weights = _initialize_weights |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/stablelm_attn_hijack_flash.py | modeling_stablelm.Attention.forward = flashattn_attn |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/stablelm_attn_hijack_flash.py | modeling_stablelm.DecoderLayer.forward = decoder_layer_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/stablelm_attn_hijack_flash.py | modeling_stablelm.StableLMEpochModel.forward = stablelm_model_forward |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/transformers_fa_utils.py | transformers.modeling_flash_attention_utils.fa_peft_integration_check = partial( |
| axolotl-ai-cloud/axolotl/src/axolotl/monkeypatch/transformers_num_parameters.py | PreTrainedModel.resize_token_embeddings = resize_token_embeddings # type: ignore[method-assign] |
| deepspeedai/DeepSpeed/deepspeed/inference/engine.py | DeepSpeedSelfAttention.num_layers = 0 |
| deepspeedai/DeepSpeed/deepspeed/linear/context_manager.py | transformers.AutoModelForCausalLM.from_config = from_config |
| deepspeedai/DeepSpeed/deepspeed/linear/context_manager.py | transformers.AutoModelForCausalLM.from_config = self._orig_causallm_config |
| deepspeedai/DeepSpeed/deepspeed/linear/context_manager.py | transformers.AutoModelForCausalLM.from_pretrained = from_pretrained |
| deepspeedai/DeepSpeed/deepspeed/linear/context_manager.py | transformers.AutoModelForCausalLM.from_pretrained = self._orig_causallm_pretrained |
| deepspeedai/DeepSpeed/deepspeed/ops/transformer/inference/ds_attention.py | DeepSpeedSelfAttention._qkv_buffers = [ |
| deepspeedai/DeepSpeed/deepspeed/ops/transformer/inference/ds_attention.py | DeepSpeedSelfAttention.num_layers = DeepSpeedSelfAttention.num_layers + 1 |
| deepspeedai/DeepSpeed/deepspeed/ops/transformer/inference/ds_mlp.py | DeepSpeedMLP._inter_w_buffers = [ |
| deepspeedai/DeepSpeed/deepspeed/ops/transformer/inference/triton/attention.py | TritonSelfAttention.num_layers = TritonSelfAttention.num_layers + 1 |
| deepspeedai/DeepSpeed/deepspeed/runtime/sequence_parallel/ulysses_sp.py | modeling_llama.LlamaMLP.forward = mlp_forward_new |
| deepspeedai/DeepSpeed/deepspeed/runtime/sequence_parallel/ulysses_sp.py | modeling_llama.LlamaMLP.forward = tiled_mlp_forward |
| hiyouga/LlamaFactory/src/llamafactory/extras/misc.py | transformers.dynamic_module_utils.check_imports = get_relative_imports |
| hiyouga/LlamaFactory/src/llamafactory/model/model_utils/longlora.py | LlamaAttention.forward = llama_attention_forward |
| hiyouga/LlamaFactory/src/llamafactory/model/model_utils/longlora.py | LlamaFlashAttention2.forward = llama_flash_attention_2_forward |
| hiyouga/LlamaFactory/src/llamafactory/model/model_utils/longlora.py | LlamaSdpaAttention.forward = llama_sdpa_attention_forward |
| hiyouga/LlamaFactory/src/llamafactory/model/model_utils/visual.py | transformers.models.llava.modeling_llava.LlavaMultiModalProjector = LlavaMultiModalProjectorForYiVL |
| hiyouga/LlamaFactory/src/llamafactory/model/patcher.py | Qwen3_5DecoderLayer.forward = _patched_decoder_forward |
| hiyouga/LlamaFactory/src/llamafactory/model/patcher.py | Qwen3_5MoeDecoderLayer.forward = _patched_decoder_forward |
| hiyouga/LlamaFactory/src/llamafactory/model/patcher.py | modeling_qwen3_omni_moe.Qwen3OmniMoeThinkerTextSparseMoeBlock = Qwen3OmniMoeThinkerTextSparseMoeBlock |
| hpcaitech/ColossalAI/colossalai/lazy/pretrained.py | PreTrainedModel.from_pretrained = classmethod(PretrainedManager.old_from_pretrained) |
| hpcaitech/ColossalAI/colossalai/lazy/pretrained.py | PreTrainedModel.from_pretrained = new_from_pretrained |
| hpcaitech/ColossalAI/colossalai/legacy/inference/quant/smoothquant/models/base_model.py | transformers.modeling_utils._init_weights = False |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | ColoAttention._flash_kernel_dispatch = ColoAttention._flash_kernel_dispatch.load() |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | ColoAttention._flash_kernel_dispatch = FlashAttentionDaoLoader() |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | ColoAttention._kernel_dispatch_map = { |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.ATTN_DONE = torch.cuda.Event() |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.CU_SEQLENS = cu_seqlens |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.HALF_INDICES = (half_idx_front, half_idx_back) |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.INNER_RING_GROUP = inner_ring_group |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.INTER_RING_GROUP = inter_ring_group |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.SP_GROUP = sp_group |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.SP_STREAM = torch.cuda.Stream() |
| hpcaitech/ColossalAI/colossalai/shardformer/layer/attn.py | RingAttention.TOTAL_SEQLEN = b * sq |
| huggingface/diffusers/examples/research_projects/pytorch_xla/inference/flux/flux_inference.py | FlashAttention.DEFAULT_BLOCK_SIZES = { |
| huggingface/diffusers/examples/research_projects/pytorch_xla/inference/flux/flux_inference_spmd.py | FlashAttention.DEFAULT_BLOCK_SIZES = { |
| huggingface/diffusers/src/diffusers/pipelines/prx/__init__.py | transformers.T5GemmaEncoder = T5GemmaEncoder |
| huggingface/trl/trl/_compat.py | transformers.cache_utils.HybridCache = Cache |
| intel/auto-round/auto_round/modeling/unfused_moe/__init__.py | transformers.modeling_utils.get_checkpoint_conversion_mapping = get_checkpoint_conversion_mapping_ar |
| intel/auto-round/auto_round/utils/common.py | transformers.modeling_utils.no_init_weights = no_init_weights |
| kvcache-ai/ktransformers/archive/kt-sft/ktransformers/operators/dynamic_attention.py | DynamicScaledDotProductAttention.cpu_infer = CPUInfer(threads_num) |
| kvcache-ai/ktransformers/archive/kt-sft/ktransformers/operators/models.py | KLlamaModel.dynamic_sdpa = DynamicScaledDotProductAttention( |
| kvcache-ai/ktransformers/archive/kt-sft/ktransformers/sft/metrics_utils/misc.py | transformers.dynamic_module_utils.check_imports = get_relative_imports |
| kvcache-ai/ktransformers/archive/ktransformers/operators/dynamic_attention.py | DynamicScaledDotProductAttention.cpu_infer = CPUInfer(threads_num) |
| kvcache-ai/ktransformers/archive/ktransformers/operators/models.py | KLlamaModel.dynamic_sdpa = DynamicScaledDotProductAttention( |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/llama4_rope.py | modeling_module.apply_rotary_emb = liger_llama4_text_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/llama4_rope.py | modeling_module.vision_apply_rotary_emb = liger_llama4_vision_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | Gemma4TextForCausalLM.forward = causal_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v2.DeepseekV2Experts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v2.DeepseekV2ForCausalLM.forward = deepseek_v2_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v2.DeepseekV2MLP = LigerQwen3MoeSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v2.DeepseekV2RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v3.DeepseekV3Experts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v3.DeepseekV3ForCausalLM.forward = deepseek_v3_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v3.DeepseekV3MLP = LigerQwen3MoeSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v3.DeepseekV3RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v4.DeepseekV4ForCausalLM.forward = deepseek_v4_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_deepseek_v4.DeepseekV4RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_exaone4.Exaone4ForCausalLM.forward = exaone4_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_exaone4.Exaone4MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_exaone4.Exaone4RMSNorm = Exaone4LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_exaone4.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_falcon_h1.FalconH1ForCausalLM.forward = falcon_h1_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_falcon_h1.FalconH1MLP = LigerFalconH1SwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_falcon_h1.FalconH1RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_falcon_h1.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma.GemmaForCausalLM.forward = gemma_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma.GemmaMLP = LigerGEGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma.GemmaRMSNorm = LigerRMSNormForGemma |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma2.Gemma2ForCausalLM.forward = gemma2_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma2.Gemma2MLP = LigerGEGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma2.Gemma2RMSNorm = LigerRMSNormForGemma2 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma2.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma3.Gemma3ForCausalLM.forward = causal_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma3.Gemma3ForConditionalGeneration.forward = multimodal_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma3.Gemma3MLP = LigerGEGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma3.Gemma3RMSNorm = LigerRMSNormForGemma3 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma3.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma3.nn.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma4.Gemma4ForCausalLM.forward = causal_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma4.Gemma4ForConditionalGeneration.forward = multimodal_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma4.Gemma4RMSNorm = LigerRMSNormForGemma4 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gemma4.Gemma4TextMLP = LigerGEGLUMLPForGemma4 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4.Glm4ForCausalLM.forward = glm4_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4.Glm4MLP = LigerPhi3SwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4.Glm4RMSNorm = LigerRMSNormForGlm4 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4v.Glm4vForConditionalGeneration.forward = glm4v_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4v.Glm4vRMSNorm = LigerRMSNormForGlm4 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4v_moe.Glm4vMoeForConditionalGeneration.forward = glm4v_moe_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4v_moe.Glm4vMoeRMSNorm = LigerRMSNormForGlm4 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4v_moe.Glm4vMoeTextNaiveMoe = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_glm4v_moe.Glm4vMoeTextRMSNorm = LigerRMSNormForGlm4 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gpt_oss.GptOssForCausalLM.forward = gpt_oss_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gpt_oss.GptOssRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_gpt_oss.apply_rotary_pos_emb = liger_rotary_pos_emb_gpt_oss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_granite.GraniteMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_granite.GraniteRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_granite.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_dense.HunYuanDenseV1ForCausalLM.forward = hunyuan_v1_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_dense.HunYuanDenseV1MLP = LigerHunyuanV1SwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_dense.HunYuanDenseV1RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_dense.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_moe.HunYuanMoEV1Experts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_moe.HunYuanMoEV1ForCausalLM.forward = hunyuan_v1_moe_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_moe.HunYuanMoEV1RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_hunyuan_v1_moe.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_internvl.InternVLForConditionalGeneration.forward = internvl_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_internvl.InternVLVisionRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_internvl.nn.LayerNorm = LigerLayerNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama.LlamaForCausalLM.forward = llama_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama.LlamaMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama.LlamaRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama4.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama4.Llama4ForCausalLM.forward = llama4_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama4.Llama4TextMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llama4.Llama4TextRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_llava.LlavaForConditionalGeneration.forward = llava_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_ministral.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_ministral.MinistralForCausalLM.forward = ministral_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_ministral.MinistralMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_ministral.MinistralRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_ministral.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mistral.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mistral.MistralForCausalLM.forward = mistral_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mistral.MistralMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mistral.MistralRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mistral.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mixtral.MixtralExperts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mixtral.MixtralForCausalLM.forward = mixtral_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mixtral.MixtralRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mixtral.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mllama.MllamaForCausalLM.forward = mllama_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mllama.MllamaTextMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mllama.MllamaTextRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mllama.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_mllama.nn.LayerNorm = LigerLayerNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_muse_glimmer.MuseGlimmerForConditionalGeneration.forward = muse_glimmer_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_muse_glimmer.MuseGlimmerRMSNorm = LigerRMSNormForMuseGlimmer |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_muse_glimmer.MuseGlimmerTextCenteredRMSNorm = LigerRMSNormForMuseGlimmerTextCentered |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_muse_glimmer.MuseGlimmerTextMLP = LigerSwiGLUMLPForMuseGlimmer |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_muse_glimmer.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_nemotron.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_nemotron.NemotronForCausalLM.forward = nemotron_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo2.Olmo2ForCausalLM.forward = olmo2_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo2.Olmo2MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo2.Olmo2RMSNorm = LigerRMSNormForOlmo2 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo2.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo3.Olmo3ForCausalLM.forward = olmo3_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo3.Olmo3MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo3.Olmo3RMSNorm = LigerRMSNormForOlmo2 # same as olmo2 |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_olmo3.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_paligemma.PaliGemmaForConditionalGeneration.forward = lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_paligemma.nn.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_phi3.Phi3ForCausalLM.forward = phi3_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_phi3.Phi3MLP = LigerPhi3SwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_phi3.Phi3RMSNorm = LigerRMSNorm # Same as Llama |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_phi3.apply_rotary_pos_emb = liger_rotary_pos_emb # Same as Gemma |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_pixtral.PixtralMLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_pixtral.PixtralRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_pixtral.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2.Qwen2ForCausalLM.forward = qwen2_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2.Qwen2MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2.Qwen2RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_5_vl.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_5_vl.Qwen2MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_5_vl.Qwen2RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_5_vl.Qwen2_5_VLForConditionalGeneration.forward = qwen2_5_vl_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_5_vl.Qwen2_5_VLRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_5_vl.apply_multimodal_rotary_pos_emb = liger_multimodal_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.CrossEntropyLoss = LigerCrossEntropyLoss |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.LayerNorm = LigerLayerNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.Qwen2MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.Qwen2RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.Qwen2VLForConditionalGeneration.forward = qwen2_vl_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.Qwen2VLRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen2_vl.apply_multimodal_rotary_pos_emb = liger_multimodal_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3.Qwen3ForCausalLM.forward = qwen3_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3.Qwen3MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3.Qwen3RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5.Qwen3_5ForCausalLM.forward = qwen3_5_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5.Qwen3_5ForConditionalGeneration.forward = qwen3_5_lce_forward_for_multimodal |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5.Qwen3_5MLP = LigerQwen3MoeSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5.Qwen3_5RMSNorm = LigerRMSNormForQwen3Next |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5.apply_rotary_pos_emb_vision = liger_rotary_pos_emb_vision |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeExperts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeForCausalLM.forward = qwen3_5_moe_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeForConditionalGeneration.forward = ( |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeRMSNorm = LigerRMSNormForQwen3Next |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5_moe.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_5_moe.apply_rotary_pos_emb_vision = liger_rotary_pos_emb_vision |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_moe.Qwen3MoeExperts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_moe.Qwen3MoeForCausalLM.forward = qwen3_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_moe.Qwen3MoeRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_moe.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_next.Qwen3NextExperts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_next.Qwen3NextForCausalLM.forward = qwen3_next_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_next.Qwen3NextRMSNorm = LigerRMSNormForQwen3Next |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_next.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl.Qwen3VLForConditionalGeneration.forward = qwen3_vl_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl.Qwen3VLTextRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl.apply_rotary_pos_emb_vision = liger_rotary_pos_emb_vision |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl_moe.Qwen3VLMoeForConditionalGeneration.forward = qwen3_vl_moe_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl_moe.Qwen3VLMoeTextExperts = LigerExperts |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl_moe.Qwen3VLMoeTextRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl_moe.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_qwen3_vl_moe.apply_rotary_pos_emb_vision = liger_rotary_pos_emb_vision |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_siglip.nn.LayerNorm = LigerLayerNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smollm3.SmolLM3ForCausalLM.forward = smollm3_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smollm3.SmolLM3MLP = LigerSwiGLUMLP |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smollm3.SmolLM3RMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smollm3.apply_rotary_pos_emb = liger_rotary_pos_emb |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smolvlm.SmolVLMForConditionalGeneration.forward = smolvlm_lce_forward |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smolvlm.SmolVLMRMSNorm = LigerRMSNorm |
| linkedin/Liger-Kernel/src/liger_kernel/transformers/monkey_patch.py | modeling_smolvlm.nn.LayerNorm = LigerLayerNorm |
| modelscope/ms-swift/swift/megatron/utils/convert_utils.py | TEDotProductAttention.forward = forward |
| modelscope/ms-swift/swift/megatron/utils/convert_utils.py | TEDotProductAttention.forward = new_forward |
| modelscope/ms-swift/swift/model/models/deepseek.py | UnlimitedOCRModel._swift_multi_gpu_patched = True |
| modelscope/ms-swift/swift/model/models/deepseek.py | UnlimitedOCRModel.forward = _patched_forward |
| modelscope/ms-swift/swift/model/models/deepseek.py | modeling_module.torch.cat = _orig_cat |
| modelscope/ms-swift/swift/model/models/deepseek.py | modeling_module.torch.cat = _safe_cat |
| modelscope/ms-swift/swift/model/models/deepseek.py | transformers.models.llama.modeling_llama.LlamaFlashAttention2 = None |
| modelscope/ms-swift/swift/model/models/internlm.py | PreTrainedModel._old_enable_input_require_grads = old_enable_input_require_grads |
| modelscope/ms-swift/swift/model/models/internlm.py | PreTrainedModel.enable_input_require_grads = patched_enable_input_require_grads |
| modelscope/ms-swift/swift/model/models/llava.py | LlavaLlamaForCausalLM.__old_forward = forward |
| modelscope/ms-swift/swift/model/models/llava.py | LlavaLlamaForCausalLM.forward = _new_forward |
| modelscope/ms-swift/swift/model/models/spark.py | modeling_module._swift_patched = True |
| modelscope/ms-swift/swift/model/models/spark.py | setattr(modeling_module, name, Spark2_5Loader._compat_mask_fn(getattr(modeling_module, name))) |
| modelscope/ms-swift/swift/model/models/valley.py | ValleyQwen2ForCausalLM._origin_forward = forward |
| modelscope/ms-swift/swift/model/models/valley.py | ValleyQwen2ForCausalLM.forward = new_forward |
| modelscope/ms-swift/swift/model/models/xing4_0.py | modeling_module.Xing4_0HyperConnection.forward = torch.compile( |
| modelscope/ms-swift/swift/model/models/xing4_0.py | modeling_module.Xing4_0PreTrainedModel._can_set_experts_implementation_cached_value = True |
| modelscope/ms-swift/swift/model/models/xing4_0.py | modeling_module._swift_patch_options = patch_options |
| modelscope/ms-swift/swift/model/patcher.py | PreTrainedModel._get_no_split_modules = _get_no_split_modules |
| modelscope/ms-swift/swift/model/patcher.py | PreTrainedModel._get_no_split_modules = _new_get_no_split_modules |
| modelscope/ms-swift/swift/model/patcher.py | PreTrainedModel.from_pretrained = _new_from_pretrained |
| modelscope/ms-swift/swift/model/patcher.py | PreTrainedModel.from_pretrained = classmethod(from_pretrained) |
| modelscope/ms-swift/swift/model/patcher.py | transformers.modeling_utils.get_balanced_memory = lambda *args, **kwargs: {} |
| modelscope/ms-swift/swift/model/patcher.py | transformers.modeling_utils.infer_auto_device_map = _infer_auto_device_map_patch |
| modelscope/ms-swift/swift/sequence_parallel/sequence_parallel.py | modeling_flash_attention_utils._flash_attention_forward = flash_attention_forward |
| modelscope/ms-swift/swift/sequence_parallel/sequence_parallel.py | modeling_flash_attention_utils._flash_attention_forward_origin = _flash_attention_forward |
| modelscope/ms-swift/swift/template/base.py | modeling_module._flash_attention_forward = _flash_attention_forward |
| modelscope/ms-swift/swift/template/base.py | modeling_module._flash_attention_forward = _origin_flash_attention_forward |
| modelscope/ms-swift/swift/template/templates/glm.py | modeling_module.create_causal_mask = new_create_causal_mask |
| modelscope/ms-swift/swift/trainers/mixin.py | transformers.modeling_utils.checkpoint = _new_checkpoint |
| modelscope/ms-swift/swift/tuners/peft.py | AdaLoraModel._create_and_replace = _create_and_replace_hook |
| modelscope/ms-swift/swift/tuners/peft.py | AdaLoraModel._create_and_replace_origin = AdaLoraModel._create_and_replace |
| modelscope/ms-swift/swift/tuners/peft.py | AdaLoraModel.forward = adalora_forward |
| modelscope/ms-swift/swift/tuners/peft.py | BOFTModel._create_and_replace = _create_and_replace_hook |
| modelscope/ms-swift/swift/tuners/peft.py | BOFTModel._create_and_replace_origin = BOFTModel._create_and_replace |
| modelscope/ms-swift/swift/tuners/peft.py | BoneModel._create_and_replace = _create_and_replace_hook |
| modelscope/ms-swift/swift/tuners/peft.py | BoneModel._create_and_replace_origin = BoneModel._create_and_replace |
| modelscope/ms-swift/swift/tuners/peft.py | FourierFTModel._create_and_replace = _create_and_replace_hook |
| modelscope/ms-swift/swift/tuners/peft.py | FourierFTModel._create_and_replace_origin = FourierFTModel._create_and_replace |
| modelscope/ms-swift/swift/tuners/peft.py | LoraModel.__init__ = __new_init__ |
| modelscope/ms-swift/swift/tuners/peft.py | LoraModel.__init_origin__ = LoraModel.__init__ |
| modelscope/ms-swift/swift/tuners/peft.py | LoraModel._cast_adapter_dtype = _cast_adapter_dtype_hook |
| modelscope/ms-swift/swift/tuners/peft.py | LoraModel._cast_adapter_dtype_origin = LoraModel._cast_adapter_dtype |
| modelscope/ms-swift/swift/tuners/peft.py | LoraModel._create_and_replace = _create_and_replace_hook |
| modelscope/ms-swift/swift/tuners/peft.py | LoraModel._create_and_replace_origin = LoraModel._create_and_replace |
| modelscope/ms-swift/swift/tuners/peft.py | PeftModel.activate_adapter = PeftModel.set_adapter |
| modelscope/ms-swift/swift/tuners/peft.py | PeftModel.create_optimizer_param_groups = create_optimizer_param_groups |
| modelscope/ms-swift/swift/tuners/peft.py | PeftModel.deactivate_adapter = partial(dummy_function, func='deactivate_adapter') |
| modelscope/ms-swift/swift/tuners/peft.py | PeftModel.load_adapter = load_adapter |
| modelscope/ms-swift/swift/tuners/peft.py | PeftModel.load_adapter_origin = PeftModel.load_adapter |
| modelscope/ms-swift/swift/tuners/peft.py | PeftModel.set_active_adapters = partial(dummy_function, func='set_active_adapters') |
| modelscope/ms-swift/swift/tuners/peft.py | VeraModel._create_and_replace = _create_and_replace_hook |
| modelscope/ms-swift/swift/tuners/peft.py | VeraModel._create_and_replace_origin = VeraModel._create_and_replace |
| sgl-project/sglang/python/sglang/multimodal_gen/runtime/layers/attention/layer.py | USPAttention._usp_a2a_stream = torch.get_device_module().Stream() |
| sgl-project/sglang/python/sglang/multimodal_gen/runtime/loader/utils.py | PreTrainedModel.post_init = lambda self: None |
| sgl-project/sglang/python/sglang/srt/multimodal/processors/qwen_vl.py | transformers.models.qwen2_vl.image_processing_qwen2_vl_fast.Qwen2VLImageProcessorFast._preprocess = fast_preprocess_cpu |
| sgl-project/sglang/python/sglang/srt/utils/hf_transformers_patches.py | modeling_llama.LlamaFlashAttention2 = modeling_llama.LlamaAttention |
| unslothai/unsloth-zoo/unsloth_zoo/compiler.py | modeling_file.__UNSLOTH_PATCHED__ = True |
| unslothai/unsloth-zoo/unsloth_zoo/compiler.py | modeling_file.__UNSLOTH_SUPPORTS_SDPA__ = False |
| unslothai/unsloth-zoo/unsloth_zoo/compiler.py | modeling_file.__UNSLOTH_SUPPORTS_SDPA__ = final_supports_sdpa |
| unslothai/unsloth-zoo/unsloth_zoo/compiler.py | modeling_file.unsloth_mhc_projection = unsloth_mhc_projection |
| unslothai/unsloth-zoo/unsloth_zoo/compiler.py | modeling_file.unsloth_sinkhorn_knopp = unsloth_sinkhorn_knopp |
| unslothai/unsloth-zoo/unsloth_zoo/compiler.py | setattr(modeling_file, key, value) |
| unslothai/unsloth-zoo/unsloth_zoo/empty_model.py | BitsAndBytesModelLoader._stack_quantization_states = ( |
| unslothai/unsloth-zoo/unsloth_zoo/empty_model.py | Gemma4ForCausalLM.get_expert_mapping = get_expert_mapping |
| unslothai/unsloth-zoo/unsloth_zoo/empty_model.py | SiglipVisionModel._init_weights = SiglipVisionModel._original_initialize_weights |
| unslothai/unsloth-zoo/unsloth_zoo/empty_model.py | SiglipVisionModel._init_weights = _init_weights |
| unslothai/unsloth-zoo/unsloth_zoo/empty_model.py | SiglipVisionModel._original_initialize_weights = SiglipVisionModel._init_weights |
| unslothai/unsloth-zoo/unsloth_zoo/gradient_checkpointing.py | transformers.modeling_utils._unsloth_old_checkpoint = current |
| unslothai/unsloth-zoo/unsloth_zoo/gradient_checkpointing.py | transformers.modeling_utils.checkpoint = ( |
| unslothai/unsloth-zoo/unsloth_zoo/gradient_checkpointing.py | transformers.modeling_utils.checkpoint = getattr( |
| unslothai/unsloth-zoo/unsloth_zoo/gradient_checkpointing.py | transformers.modeling_utils.checkpoint = shim |
| unslothai/unsloth-zoo/unsloth_zoo/gradient_checkpointing.py | transformers.modeling_utils.checkpoint = unsloth_checkpoint |
| unslothai/unsloth-zoo/unsloth_zoo/loss_utils.py | UnslothForCausalLMLoss._unsloth_fused_loss = UnslothForCausalLMLoss |
| unslothai/unsloth-zoo/unsloth_zoo/loss_utils.py | transformers.modeling_utils.PreTrainedModel.loss_function = \ |
| unslothai/unsloth-zoo/unsloth_zoo/mlx/loader.py | modeling_llama.LlamaFlashAttention2 = modeling_llama.LlamaAttention |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/common.py | setattr(modeling_module, name, value) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/compiled_model_identity.py | PreTrainedModel.is_custom_code = method |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/deepseek_v3_moe.py | DeepseekV3ForCausalLM.__UNSLOTH_SUPPORTS_RETURN_HIDDEN_STATES__ = True |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/deepseek_v3_moe.py | DeepseekV3ForCausalLM.forward = _patched_causal_lm_forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gemma3n.py | Gemma3nTextAttention.forward = forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gemma4.py | Gemma4AudioAttention.forward = forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gemma4.py | Gemma4TextAttention.forward = _make_gemma4_attention_carrier_forward(Gemma4TextAttention.forward) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gemma4.py | LoraModel._create_and_replace = create_and_replace |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | GptOssForCausalLM._unsloth_grpo_patched = True |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | GptOssForCausalLM.forward = _patched_causal_lm_forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | GptOssModel.forward = forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | GptOssPreTrainedModel._unsloth_init_weights_fixed = True |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.generation.utils.create_masks_for_generate = wrap(transformers.generation.utils.create_masks_for_generate) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.masking_utils.__patched_causal_mask__ = True |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.masking_utils._old_create_causal_mask = compile_mask_builder(transformers.masking_utils.create_causal_mask) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.masking_utils._old_create_sliding_window_causal_mask = compile_mask_builder(transformers.masking_utils.create_sliding_window_causal_mask) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.masking_utils.create_causal_mask = wrap(create_causal_mask) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.masking_utils.create_masks_for_generate = wrap(transformers.masking_utils.create_masks_for_generate) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.masking_utils.create_sliding_window_causal_mask = wrap(create_sliding_window_causal_mask) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.GptOssExperts = GptOssExpertsBnb4bit |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.GptOssExperts = \ |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.GptOssExperts.forward = experts_forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.GptOssTopKRouter = GptOssTopKRouter |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.GptOssTopKRouter = \ |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss._original_GptOssExperts = \ |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss._original_GptOssTopKRouter = \ |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.create_causal_mask = transformers.masking_utils.create_causal_mask |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.models.gpt_oss.modeling_gpt_oss.create_sliding_window_causal_mask = transformers.masking_utils.create_sliding_window_causal_mask |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.quantizers.quantizer_mxfp4.Mxfp4HfQuantizer._lazy_import_kernels = lambda *args, **kwargs: triton_kernels |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.quantizers.quantizer_mxfp4.Mxfp4HfQuantizer.is_trainable = lambda *args, **kwargs: True |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.quantizers.quantizer_mxfp4.is_kernels_available = is_kernels_available |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/gpt_oss.py | transformers.quantizers.quantizer_mxfp4.is_kernels_available = lambda *args, **kwargs: False |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/misc.py | HiggsAudioV2Embeddings.forward = forward |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/misc.py | ModernBertModel._update_attention_mask = _update_attention_mask_contiguous |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/misc.py | transformers.models.mllama.modeling_mllama.MllamaVisionEncoderLayer = MllamaVisionEncoderLayer |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/misc.py | transformers.utils.import_utils.is_causal_conv1d_available = lambda: False |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/moe_utils.py | PeftModel.load_adapter = patched |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/moe_utils.py | PeftModel.save_pretrained = patched |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/moe_utils_bnb4bit.py | modeling_utils.get_model_conversion_mapping = patched_get_model_conversion_mapping |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/mxfp4.py | PreTrainedModel.save_pretrained = save_pretrained |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/mxfp4.py | modeling_utils._get_device_map = _get_device_map |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/mxfp4.py | transformers.integrations.mxfp4.convert_moe_packed_tensors_cpu = convert_moe_packed_tensors_cpu |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/qwen3_5_moe.py | transformers.models.qwen3_5_moe.modeling_qwen3_5_moe.Qwen3_5MoeExperts._unsloth_lora_extractor_fn = staticmethod(_qwen3_5_lora_extractor) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/qwen3_moe.py | transformers.models.qwen3_moe.modeling_qwen3_moe.Qwen3MoeExperts._unsloth_lora_extractor_fn = staticmethod(_qwen3_lora_extractor) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/qwen3_next_moe.py | transformers.models.qwen3_next.modeling_qwen3_next.Qwen3NextExperts._unsloth_lora_extractor_fn = staticmethod(_qwen3_next_lora_extractor) |
| unslothai/unsloth-zoo/unsloth_zoo/temporary_patches/qwen3_vl_moe.py | transformers.models.qwen3_vl_moe.modeling_qwen3_vl_moe.Qwen3VLMoeTextExperts._unsloth_lora_extractor_fn = staticmethod(_qwen3_vl_lora_extractor) |
| unslothai/unsloth-zoo/unsloth_zoo/vllm_utils.py | Qwen3MoeForCausalLM.is_3d_moe_weight = True |
| unslothai/unsloth/studio/backend/core/inference/safetensors_pread.py | modeling_utils.safe_open = _original |
| unslothai/unsloth/studio/backend/core/inference/safetensors_pread.py | modeling_utils.safe_open = _with_pread(current) |
| unslothai/unsloth/unsloth/import_fixes.py | PreTrainedModel.disable_input_require_grads = _patched_single_disable_input_require_grads |
| unslothai/unsloth/unsloth/import_fixes.py | PreTrainedModel.enable_input_require_grads = _patched_enable_input_require_grads |
| unslothai/unsloth/unsloth/import_fixes.py | PreTrainedModel.enable_input_require_grads = _patched_single_enable_input_require_grads |
| unslothai/unsloth/unsloth/import_fixes.py | PreTrainedModel.post_init = post_init |
| unslothai/unsloth/unsloth/import_fixes.py | modeling_utils.revert_weight_conversion = revert_weight_conversion |
| unslothai/unsloth/unsloth/kernels/rms_layernorm.py | transformers.models.llama.modeling_llama.LlamaRMSNorm = LlamaRMSNorm |
| unslothai/unsloth/unsloth/kernels/rms_layernorm.py | transformers.models.llama.modeling_llama.LlamaRMSNorm = Unsloth_LlamaRMSNorm |
| unslothai/unsloth/unsloth/kernels/rms_layernorm.py | transformers.models.mllama.modeling_mllama.MllamaTextRMSNorm = MllamaTextRMSNorm |
| unslothai/unsloth/unsloth/kernels/rms_layernorm.py | transformers.models.mllama.modeling_mllama.MllamaTextRMSNorm = Unsloth_MllamaTextRMSNorm |
| unslothai/unsloth/unsloth/models/_utils.py | PreTrainedModel._get_key_renaming_mapping = _get_key_renaming_mapping |
| unslothai/unsloth/unsloth/models/_utils.py | PreTrainedModel._get_key_renaming_mapping = original |
| unslothai/unsloth/unsloth/models/_utils.py | modeling_utils.get_model_conversion_mapping = get_model_conversion_mapping |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.integrations.bitsandbytes.should_convert_module = patched_should_convert_module |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.trainer.get_model_param_count = get_model_param_count |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.trainer_pt_utils.get_model_param_count = get_model_param_count |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.utils.import_utils.is_flash_attn_2_available = ( |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.utils.import_utils.is_flash_attn_2_available = lambda *args, **kwargs: ( |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.utils.import_utils.is_flash_attn_2_available = lambda *args, **kwargs: False |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.utils.is_flash_attn_2_available = lambda *args, **kwargs: False |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.utils.is_openai_available = _lazy_openai_import_check( |
| unslothai/unsloth/unsloth/models/_utils.py | transformers.utils.quantization_config.BitsAndBytesConfig.__init__ = _BitsAndBytesConfig__init__ |
| unslothai/unsloth/unsloth/models/cohere.py | CohereAttention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/cohere.py | CohereAttention.forward = CohereAttention_fast_forward |
| unslothai/unsloth/unsloth/models/cohere.py | CohereDecoderLayer.forward = CohereDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/cohere.py | CohereFlashAttention2.forward = CohereAttention_fast_forward |
| unslothai/unsloth/unsloth/models/cohere.py | CohereForCausalLM.forward = CausalLM_fast_forward(CohereModel_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/cohere.py | CohereModel.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/cohere.py | CohereSdpaAttention.forward = CohereAttention_fast_forward |
| unslothai/unsloth/unsloth/models/cohere.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/cohere.py | transformers.models.cohere.modeling_cohere.CohereRotaryEmbedding = LlamaRotaryEmbedding |
| unslothai/unsloth/unsloth/models/compressed_tensors_int4.py | PeftModel.save_pretrained = save_pretrained |
| unslothai/unsloth/unsloth/models/falcon_h1.py | FalconH1Attention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/falcon_h1.py | FalconH1Attention.forward = FalconH1Attention_fast_forward |
| unslothai/unsloth/unsloth/models/falcon_h1.py | FalconH1DecoderLayer.forward = FalconH1DecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/falcon_h1.py | FalconH1ForCausalLM.forward = CausalLM_fast_forward( |
| unslothai/unsloth/unsloth/models/falcon_h1.py | FalconH1Model.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/falcon_h1.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/falcon_h1.py | transformers.models.falcon_h1.modeling_falcon_h1.FalconH1RMSNorm = Unsloth_FalconH1RMSNorm |
| unslothai/unsloth/unsloth/models/falcon_h1.py | transformers.models.falcon_h1.modeling_falcon_h1.FalconH1RotaryEmbedding = ( |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaAttention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaAttention.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaDecoderLayer.forward = GemmaDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaFlashAttention2.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaForCausalLM.forward = CausalLM_fast_forward(GemmaModel_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaModel.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/gemma.py | GemmaSdpaAttention.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/gemma.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/gemma.py | transformers.models.gemma.modeling_gemma.GemmaRotaryEmbedding = GemmaFixedRotaryEmbedding |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2Attention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2Attention.forward = Gemma2Attention_fast_forward |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2DecoderLayer.forward = Gemma2DecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2FlashAttention2.forward = Gemma2Attention_fast_forward |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2ForCausalLM.forward = CausalLM_fast_forward(Gemma2Model_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2Model.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/gemma2.py | Gemma2SdpaAttention.forward = Gemma2Attention_fast_forward |
| unslothai/unsloth/unsloth/models/gemma2.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/gemma2.py | transformers.models.gemma2.modeling_gemma2.Gemma2RotaryEmbedding = GemmaFixedRotaryEmbedding |
| unslothai/unsloth/unsloth/models/granite.py | GraniteAttention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/granite.py | GraniteAttention.forward = GraniteAttention_fast_forward |
| unslothai/unsloth/unsloth/models/granite.py | GraniteDecoderLayer.forward = GraniteDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/granite.py | GraniteFlashAttention2.forward = GraniteAttention_fast_forward |
| unslothai/unsloth/unsloth/models/granite.py | GraniteForCausalLM.__init__ = patched_init(GraniteForCausalLM.__init__) |
| unslothai/unsloth/unsloth/models/granite.py | GraniteForCausalLM.forward = CausalLM_fast_forward(GraniteModel_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/granite.py | GraniteModel.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/granite.py | GraniteSdpaAttention.forward = GraniteAttention_fast_forward |
| unslothai/unsloth/unsloth/models/granite.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/granite.py | transformers.models.granite.modeling_granite.GraniteRotaryEmbedding = GraniteRotaryEmbedding |
| unslothai/unsloth/unsloth/models/llama.py | FastLlamaModel.from_pretrained = staticmethod( |
| unslothai/unsloth/unsloth/models/llama.py | FastLlamaModel.get_peft_model = staticmethod( |
| unslothai/unsloth/unsloth/models/llama.py | LlamaAttention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/llama.py | LlamaAttention.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/llama.py | LlamaDecoderLayer.forward = LlamaDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/llama.py | LlamaFlashAttention2.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/llama.py | LlamaForCausalLM.forward = CausalLM_fast_forward(LlamaModel_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/llama.py | LlamaModel.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/llama.py | LlamaModel_fast_forward_inference_custom._unsloth_graph_attention = ( |
| unslothai/unsloth/unsloth/models/llama.py | LlamaSdpaAttention.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/llama.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/llama.py | transformers.models.llama.modeling_llama.LlamaLinearScalingRotaryEmbedding = ( |
| unslothai/unsloth/unsloth/models/llama.py | transformers.models.llama.modeling_llama.LlamaRotaryEmbedding = LlamaRotaryEmbedding |
| unslothai/unsloth/unsloth/models/loader.py | PreTrainedModel.from_pretrained = classmethod(wrapped) |
| unslothai/unsloth/unsloth/models/mistral.py | MistralAttention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/mistral.py | MistralAttention.forward = MistralAttention_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | MistralDecoderLayer.forward = LlamaDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | MistralFlashAttention2.forward = MistralAttention_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | MistralForCausalLM.forward = MistralForCausalLM_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | MistralModel.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | MistralSdpaAttention.forward = MistralAttention_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/mistral.py | transformers.models.mistral.modeling_mistral.MistralRotaryEmbedding = LlamaRotaryEmbedding |
| unslothai/unsloth/unsloth/models/qwen2.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2Attention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2Attention.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2DecoderLayer.forward = LlamaDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2FlashAttention2.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2ForCausalLM.forward = CausalLM_fast_forward(LlamaModel_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2Model.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/qwen2.py | Qwen2SdpaAttention.forward = LlamaAttention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen2.py | transformers.models.qwen2.modeling_qwen2.Qwen2RotaryEmbedding = LlamaRotaryEmbedding |
| unslothai/unsloth/unsloth/models/qwen3.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3Attention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3Attention.forward = Qwen3Attention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3DecoderLayer.forward = LlamaDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3FlashAttention2.forward = Qwen3Attention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3ForCausalLM.forward = CausalLM_fast_forward( |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3Model.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3.py | Qwen3SdpaAttention.forward = Qwen3Attention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3.py | transformers.models.qwen3.modeling_qwen3.Qwen3RotaryEmbedding = LlamaRotaryEmbedding |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | PeftModelForCausalLM.forward = PeftModel_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | Qwen3MoeAttention.__init__ = eval(init_name) |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | Qwen3MoeAttention.forward = Qwen3Attention_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | Qwen3MoeDecoderLayer.forward = Qwen3MoeDecoderLayer_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | Qwen3MoeForCausalLM.forward = CausalLM_fast_forward(LlamaModel_fast_forward_inference) |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | Qwen3MoeMLP.forward = fast_swiglu_inference # This is analogous to Dense models' MLP |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | Qwen3MoeModel.forward = LlamaModel_fast_forward |
| unslothai/unsloth/unsloth/models/qwen3_moe.py | transformers.models.qwen3_moe.modeling_qwen3_moe.Qwen3MoeRotaryEmbedding = ( |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | AutoModel.from_pretrained = original_model_from_pretrained |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | AutoModel.from_pretrained = return_existing_model |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | modeling_distilbert.DistilBertModel.forward = forward |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | modeling_mpnet.MPNetEncoder.forward = forward |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | modeling_mpnet.MPNetModel._set_gradient_checkpointing = _set_gradient_checkpointing |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | modeling_mpnet.MPNetModel.supports_gradient_checkpointing = True |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | transformers.AutoModelForCausalLM = original_causal |
| unslothai/unsloth/unsloth/models/sentence_transformer.py | transformers.AutoModelForCausalLM = transformers.AutoModel |
| unslothai/unsloth/unsloth/models/tp_native.py | PreTrainedModel.from_pretrained = classmethod(from_pretrained) |
| unslothai/unsloth/unsloth/models/tp_native.py | PreTrainedModel.from_pretrained = descriptor |
| unslothai/unsloth/unsloth/models/vision.py | FastBaseModel.from_pretrained = staticmethod( |
| unslothai/unsloth/unsloth/models/vision.py | FastBaseModel.get_peft_model = staticmethod( |
| unslothai/unsloth/unsloth/models/vision.py | PreTrainedModel.to = original_to |
| unslothai/unsloth/unsloth/models/vision.py | PreTrainedModel.to = to |
| verl-project/verl/verl/models/mcore/patch.py | GPTModel._postprocess = _patched |
| verl-project/verl/verl/models/mcore/patch.py | MLASelfAttention.get_query_key_value_tensors = patch_get_query_key_value_tensors |
| verl-project/verl/verl/models/mcore/patch.py | MultiLatentAttention.forward = patch_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | AutoModelForCausalLMWithValueHead.state_dict = state_dict |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Glm4vModel.forward = glm4v_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Glm4vTextAttention.forward = glm4v_attn_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen2VLAttention.forward = qwen2_vl_attn_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen2VLModel.forward = qwen2_vl_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen2_5_VLAttention.forward = qwen2_vl_attn_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen2_5_VLModel.forward = qwen2_vl_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3VLModel.forward = qwen3_vl_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3VLMoeModel.forward = qwen3_vl_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3VLMoeVisionModel.fast_pos_embed_interpolate = fast_pos_embed_interpolate |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3VLVisionModel.fast_pos_embed_interpolate = fast_pos_embed_interpolate |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3_5DecoderLayer.forward = qwen3_5_decoder_layer_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3_5Model.forward = qwen3_5_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3_5MoeDecoderLayer.forward = qwen3_5_decoder_layer_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3_5MoeModel.forward = qwen3_5_base_forward |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3_5MoeVisionModel.fast_pos_embed_interpolate = fast_pos_embed_interpolate |
| verl-project/verl/verl/models/transformers/monkey_patch.py | Qwen3_5VisionModel.fast_pos_embed_interpolate = fast_pos_embed_interpolate |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen2.Qwen2MLP.forward = silu_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen2.Qwen2RMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen2.apply_rotary_pos_emb = apply_rotary_pos_emb_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen2_5_vl.Qwen2RMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen2_5_vl.Qwen2_5_VLMLP.forward = silu_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen2_5_vl.Qwen2_5_VLRMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3.Qwen3MLP.forward = silu_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3.Qwen3RMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3.apply_rotary_pos_emb = apply_rotary_pos_emb_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5.Qwen3_5RMSNorm.forward = qwen3_next_rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5.Qwen3_5RMSNormGated.forward = qwen3_next_rms_norm_forward_gated_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5.apply_rotary_pos_emb = qwen3_next_apply_rotary_pos_emb_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeExperts.forward = qwen3_5_moe_experts_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeRMSNorm.forward = qwen3_next_rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5_moe.Qwen3_5MoeRMSNormGated.forward = qwen3_next_rms_norm_forward_gated_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_5_moe.apply_rotary_pos_emb = qwen3_next_apply_rotary_pos_emb_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_moe.Qwen3MoeRMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_moe.Qwen3MoeSparseMoeBlock.forward = qwen3_moe_sparse_moe_block_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_moe.apply_rotary_pos_emb = apply_rotary_pos_emb_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_next.Qwen3NextRMSNorm.forward = qwen3_next_rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_next.Qwen3NextRMSNormGated.forward = qwen3_next_rms_norm_forward_gated_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_next.Qwen3NextSparseMoeBlock.forward = qwen3_next_sparse_moe_block_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_next.apply_rotary_pos_emb = qwen3_next_apply_rotary_pos_emb_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_vl.Qwen3VLTextMLP.forward = silu_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_vl.Qwen3VLTextRMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_vl_moe.Qwen3VLMoeTextRMSNorm.forward = rms_norm_forward_npu |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_vl_moe.Qwen3VLMoeTextSparseMoeBlock = NPUQwen3VLMoeTextSparseMoeBlock |
| verl-project/verl/verl/models/transformers/npu_patch.py | modeling_qwen3_vl_moe.apply_rotary_pos_emb = apply_rotary_pos_emb_npu |
| vllm-project/llm-compressor/examples/quantizing_moe/deepseek_v4_example.py | DeepseekV4PreTrainedModel._keep_in_fp32_modules_strict = set() |
| vllm-project/llm-compressor/examples/quantizing_moe/deepseek_v4_mixed_mxfp4_mxfp8.py | DeepseekV4PreTrainedModel._keep_in_fp32_modules_strict = set() |
| vllm-project/llm-compressor/examples/quantizing_moe/deepseek_v4_pro_example.py | DeepseekV4PreTrainedModel._keep_in_fp32_modules_strict = set() |
| vllm-project/llm-compressor/src/llmcompressor/modeling/moe/linearize.py | transformers.integrations.finegrained_fp8.replace_with_fp8_linear = ( |
| vllm-project/vllm/vllm/model_executor/models/adapters.py | ModelForEmbedding.__name__ = _get_pooling_model_name(cls.__name__, "ForEmbedding") |
| vllm-project/vllm/vllm/model_executor/models/expert_substitution.py | ExpertSubstitutionModel.__name__ = model_cls.__name__ |
| vllm-project/vllm/vllm/model_executor/models/expert_substitution.py | ExpertSubstitutionModel.__qualname__ = model_cls.__qualname__ |
| vllm-project/vllm/vllm/v1/attention/backends/composite.py | CompositeAttentionBackend.__module__ = module |
| vllm-project/vllm/vllm/v1/attention/backends/composite.py | CompositeAttentionBackend.__name__ = name |
| vllm-project/vllm/vllm/v1/attention/backends/composite.py | CompositeAttentionBackend.__qualname__ = name |