Upload 2025-05-26/runs/609-15245141532/ci_results_run_models_gpu/new_failures_against_15245120897.json with huggingface_hub

2025-05-26/runs/609-15245141532/ci_results_run_models_gpu/new_failures_against_15245120897.json

@@ -0,0 +1,544 @@
+{
+    "blip_2": {
+        "single-gpu": [
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ForConditionalGenerationDecoderOnlyTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ForConditionalGenerationDecoderOnlyTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ModelTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextModelWithProjectionTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextModelWithProjectionTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2VisionModelWithProjectionTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2VisionModelWithProjectionTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextRetrievalModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextRetrievalModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ForConditionalGenerationDecoderOnlyTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ForConditionalGenerationDecoderOnlyTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2ModelTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextModelWithProjectionTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextModelWithProjectionTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2VisionModelWithProjectionTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2VisionModelWithProjectionTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextRetrievalModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/blip_2/test_modeling_blip_2.py::Blip2TextRetrievalModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "cohere2": {
+        "single-gpu": [
+            "tests/models/cohere2/test_modeling_cohere2.py::Cohere2IntegrationTest::test_export_static_cache",
+            "tests/models/cohere2/test_modeling_cohere2.py::Cohere2IntegrationTest::test_generation_beyond_sliding_window_0_flash_attention_2",
+            "tests/models/cohere2/test_modeling_cohere2.py::Cohere2IntegrationTest::test_generation_beyond_sliding_window_3_eager",
+            "tests/models/cohere2/test_modeling_cohere2.py::Cohere2IntegrationTest::test_model_fp16",
+            "tests/models/cohere2/test_modeling_cohere2.py::Cohere2IntegrationTest::test_model_pipeline_bf16"
+        ],
+        "multi-gpu": []
+    },
+    "conditional_detr": {
+        "single-gpu": [
+            "tests/models/conditional_detr/test_modeling_conditional_detr.py::ConditionalDetrModelIntegrationTests::test_inference_object_detection_head"
+        ],
+        "multi-gpu": [
+            "tests/models/conditional_detr/test_modeling_conditional_detr.py::ConditionalDetrModelIntegrationTests::test_inference_object_detection_head"
+        ]
+    },
+    "dac": {
+        "single-gpu": [],
+        "multi-gpu": [
+            "tests/models/dac/test_modeling_dac.py::DacIntegrationTest::test_integration_16khz"
+        ]
+    },
+    "data2vec": {
+        "single-gpu": [
+            "tests/models/data2vec/test_modeling_data2vec_vision.py::Data2VecVisionModelTest::test_sdpa_can_compile_dynamic"
+        ],
+        "multi-gpu": [
+            "tests/models/data2vec/test_modeling_data2vec_vision.py::Data2VecVisionModelTest::test_sdpa_can_compile_dynamic"
+        ]
+    },
+    "deepseek_v3": {
+        "single-gpu": [
+            "tests/models/deepseek_v3/test_modeling_deepseek_v3.py::DeepseekV3ModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/deepseek_v3/test_modeling_deepseek_v3.py::DeepseekV3ModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "deformable_detr": {
+        "single-gpu": [
+            "tests/models/deformable_detr/test_modeling_deformable_detr.py::DeformableDetrModelIntegrationTests::test_inference_object_detection_head"
+        ],
+        "multi-gpu": [
+            "tests/models/deformable_detr/test_modeling_deformable_detr.py::DeformableDetrModelIntegrationTests::test_inference_object_detection_head"
+        ]
+    },
+    "depth_pro": {
+        "single-gpu": [
+            "tests/models/depth_pro/test_modeling_depth_pro.py::DepthProModelTest::test_sdpa_can_compile_dynamic"
+        ],
+        "multi-gpu": [
+            "tests/models/depth_pro/test_modeling_depth_pro.py::DepthProModelTest::test_sdpa_can_compile_dynamic"
+        ]
+    },
+    "diffllama": {
+        "single-gpu": [
+            "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "dpr": {
+        "single-gpu": [
+            "tests/models/dpr/test_modeling_dpr.py::DPRModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/dpr/test_modeling_dpr.py::DPRModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "dpt": {
+        "single-gpu": [
+            "tests/models/dpt/test_modeling_dpt_hybrid.py::DPTModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/dpt/test_modeling_dpt_hybrid.py::DPTModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/dpt/test_modeling_dpt_hybrid.py::DPTModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/dpt/test_modeling_dpt_hybrid.py::DPTModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "emu3": {
+        "single-gpu": [
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_attention_outputs",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_batching_equivalence",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_determinism",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_feed_forward_chunking",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_hidden_states_output",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_model_outputs_equivalence",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_resize_tokens_embeddings",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_retain_grad_hidden_states_attentions",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_save_load"
+        ],
+        "multi-gpu": [
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_attention_outputs",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_batching_equivalence",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_determinism",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_feed_forward_chunking",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_hidden_states_output",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_model_outputs_equivalence",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_resize_tokens_embeddings",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_retain_grad_hidden_states_attentions",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_save_load",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_model_parallelism",
+            "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_multi_gpu_data_parallel_forward"
+        ]
+    },
+    "fastspeech2_conformer": {
+        "single-gpu": [
+            "tests/models/fastspeech2_conformer/test_modeling_fastspeech2_conformer.py::FastSpeech2ConformerModelTest::test_model_outputs_equivalence"
+        ],
+        "multi-gpu": []
+    },
+    "gemma": {
+        "single-gpu": [
+            "tests/models/gemma/test_modeling_gemma.py::GemmaModelTest::test_torch_compile_for_training",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_export_static_cache",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_model_2b_bf16_dola",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_model_7b_bf16",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_model_7b_fp16_static_cache"
+        ],
+        "multi-gpu": [
+            "tests/models/gemma/test_modeling_gemma.py::GemmaModelTest::test_torch_compile_for_training",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_export_static_cache",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_model_2b_bf16_dola",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_model_7b_bf16",
+            "tests/models/gemma/test_modeling_gemma.py::GemmaIntegrationTest::test_model_7b_fp16_static_cache"
+        ]
+    },
+    "gemma3": {
+        "single-gpu": [
+            "tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "glm4": {
+        "single-gpu": [
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_bf16",
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_eager",
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_fp16",
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_sdpa"
+        ],
+        "multi-gpu": [
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_bf16",
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_eager",
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_fp16",
+            "tests/models/glm4/test_modeling_glm4.py::Glm4IntegrationTest::test_model_9b_sdpa"
+        ]
+    },
+    "got_ocr2": {
+        "single-gpu": [
+            "tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_generate_continue_from_past_key_values"
+        ],
+        "multi-gpu": [
+            "tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_generate_continue_from_past_key_values"
+        ]
+    },
+    "gpt2": {
+        "single-gpu": [
+            "tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelLanguageGenerationTest::test_gpt2_sample"
+        ],
+        "multi-gpu": [
+            "tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelLanguageGenerationTest::test_gpt2_sample"
+        ]
+    },
+    "gpt_bigcode": {
+        "single-gpu": [
+            "tests/models/gpt_bigcode/test_modeling_gpt_bigcode.py::GPTBigCodeModelTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/gpt_bigcode/test_modeling_gpt_bigcode.py::GPTBigCodeMHAModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/gpt_bigcode/test_modeling_gpt_bigcode.py::GPTBigCodeModelTest::test_sdpa_can_dispatch_on_flash",
+            "tests/models/gpt_bigcode/test_modeling_gpt_bigcode.py::GPTBigCodeMHAModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "gpt_neox": {
+        "single-gpu": [
+            "tests/models/gpt_neox/test_modeling_gpt_neox.py::GPTNeoXLanguageGenerationTest::test_lm_generate_flex_attn_gptneox"
+        ],
+        "multi-gpu": [
+            "tests/models/gpt_neox/test_modeling_gpt_neox.py::GPTNeoXLanguageGenerationTest::test_lm_generate_flex_attn_gptneox"
+        ]
+    },
+    "helium": {
+        "single-gpu": [
+            "tests/models/helium/test_modeling_helium.py::GemmaModelTest::test_torch_compile_for_training",
+            "tests/models/helium/test_modeling_helium.py::HeliumModelTest::test_torch_compile_for_training"
+        ],
+        "multi-gpu": [
+            "tests/models/helium/test_modeling_helium.py::GemmaModelTest::test_torch_compile_for_training",
+            "tests/models/helium/test_modeling_helium.py::HeliumModelTest::test_torch_compile_for_training"
+        ]
+    },
+    "idefics2": {
+        "single-gpu": [
+            "tests/models/idefics2/test_modeling_idefics2.py::Idefics2ModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/idefics2/test_modeling_idefics2.py::Idefics2ForConditionalGenerationModelTest::test_sdpa_can_compile_dynamic"
+        ],
+        "multi-gpu": [
+            "tests/models/idefics2/test_modeling_idefics2.py::Idefics2ModelTest::test_sdpa_can_compile_dynamic",
+            "tests/models/idefics2/test_modeling_idefics2.py::Idefics2ForConditionalGenerationModelTest::test_sdpa_can_compile_dynamic"
+        ]
+    },
+    "instructblip": {
+        "single-gpu": [
+            "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_sdpa_can_compile_dynamic",
+            "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_sdpa_can_compile_dynamic",
+            "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "instructblipvideo": {
+        "single-gpu": [
+            "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_sdpa_can_compile_dynamic",
+            "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_sdpa_can_compile_dynamic",
+            "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "internvl": {
+        "single-gpu": [
+            "tests/models/internvl/test_video_processor_internvl.py::InternVLVideoProcessingTest::test_can_compile_fast_video_processor"
+        ],
+        "multi-gpu": [
+            "tests/models/internvl/test_video_processor_internvl.py::InternVLVideoProcessingTest::test_can_compile_fast_video_processor"
+        ]
+    },
+    "jamba": {
+        "single-gpu": [
+            "tests/models/jamba/test_modeling_jamba.py::JambaModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/jamba/test_modeling_jamba.py::JambaModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "jetmoe": {
+        "single-gpu": [
+            "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeModelTest::test_sdpa_equivalence",
+            "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeIntegrationTest::test_model_8b_batched_generation",
+            "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeIntegrationTest::test_model_8b_generation",
+            "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeIntegrationTest::test_model_8b_logits"
+        ],
+        "multi-gpu": []
+    },
+    "kosmos2": {
+        "single-gpu": [
+            "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "llama": {
+        "single-gpu": [
+            "tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_torch_compile_for_training",
+            "tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_llama_3_1_hard",
+            "tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits",
+            "tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits_bf16"
+        ],
+        "multi-gpu": [
+            "tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_torch_compile_for_training",
+            "tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_llama_3_1_hard",
+            "tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits",
+            "tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits_bf16"
+        ]
+    },
+    "mistral": {
+        "single-gpu": [
+            "tests/models/mistral/test_modeling_mistral.py::MistralIntegrationTest::test_model_7b_logits"
+        ],
+        "multi-gpu": [
+            "tests/models/mistral/test_modeling_mistral.py::MistralIntegrationTest::test_model_7b_logits"
+        ]
+    },
+    "mixtral": {
+        "single-gpu": [
+            "tests/models/mixtral/test_modeling_mixtral.py::MixtralIntegrationTest::test_small_model_logits",
+            "tests/models/mixtral/test_modeling_mixtral.py::MixtralIntegrationTest::test_small_model_logits_batched"
+        ],
+        "multi-gpu": [
+            "tests/models/mixtral/test_modeling_mixtral.py::MixtralIntegrationTest::test_small_model_logits",
+            "tests/models/mixtral/test_modeling_mixtral.py::MixtralIntegrationTest::test_small_model_logits_batched"
+        ]
+    },
+    "mllama": {
+        "single-gpu": [
+            "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "modernbert": {
+        "single-gpu": [
+            "tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "moonshine": {
+        "single-gpu": [],
+        "multi-gpu": [
+            "tests/models/moonshine/test_modeling_moonshine.py::MoonshineModelTest::test_sdpa_can_compile_dynamic"
+        ]
+    },
+    "nemotron": {
+        "single-gpu": [
+            "tests/models/nemotron/test_modeling_nemotron.py::NemotronModelTest::test_torch_compile_for_training"
+        ],
+        "multi-gpu": []
+    },
+    "olmo": {
+        "single-gpu": [
+            "tests/models/olmo/test_modeling_olmo.py::OlmoIntegrationTest::test_model_7b_greedy_generation",
+            "tests/models/olmo/test_modeling_olmo.py::OlmoIntegrationTest::test_model_7b_twin_2t_logits"
+        ],
+        "multi-gpu": []
+    },
+    "olmo2": {
+        "single-gpu": [
+            "tests/models/olmo2/test_modeling_olmo2.py::Olmo2IntegrationTest::test_model_7b_logits"
+        ],
+        "multi-gpu": []
+    },
+    "olmoe": {
+        "single-gpu": [
+            "tests/models/olmoe/test_modeling_olmoe.py::OlmoeIntegrationTest::test_model_7b_greedy_generation"
+        ],
+        "multi-gpu": []
+    },
+    "phi3": {
+        "single-gpu": [
+            "tests/models/phi3/test_modeling_phi3.py::Phi3IntegrationTest::test_export_static_cache"
+        ],
+        "multi-gpu": [
+            "tests/models/phi3/test_modeling_phi3.py::Phi3IntegrationTest::test_export_static_cache"
+        ]
+    },
+    "phimoe": {
+        "single-gpu": [
+            "tests/models/phimoe/test_modeling_phimoe.py::PhimoeModelTest::test_sdpa_equivalence",
+            "tests/models/phimoe/test_modeling_phimoe.py::PhimoeIntegrationTest::test_model_phimoe_instruct_logits",
+            "tests/models/phimoe/test_modeling_phimoe.py::PhimoeIntegrationTest::test_phimoe_instruct_generation"
+        ],
+        "multi-gpu": [
+            "tests/models/phimoe/test_modeling_phimoe.py::PhimoeModelTest::test_sdpa_equivalence",
+            "tests/models/phimoe/test_modeling_phimoe.py::PhimoeIntegrationTest::test_model_phimoe_instruct_logits",
+            "tests/models/phimoe/test_modeling_phimoe.py::PhimoeIntegrationTest::test_phimoe_instruct_generation"
+        ]
+    },
+    "pixtral": {
+        "single-gpu": [
+            "tests/models/pixtral/test_modeling_pixtral.py::PixtralVisionModelModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/pixtral/test_modeling_pixtral.py::PixtralVisionModelModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "qwen2_5_omni": {
+        "single-gpu": [
+            "tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_generate_continue_from_past_key_values"
+        ],
+        "multi-gpu": [
+            "tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_generate_continue_from_past_key_values"
+        ]
+    },
+    "qwen2_5_vl": {
+        "single-gpu": [],
+        "multi-gpu": [
+            "tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLModelTest::test_generate_continue_from_past_key_values"
+        ]
+    },
+    "qwen2_vl": {
+        "single-gpu": [
+            "tests/models/qwen2_vl/test_modeling_qwen2_vl.py::Qwen2VLModelTest::test_generate_continue_from_past_key_values",
+            "tests/models/qwen2_vl/test_video_processing_qwen2_vl.py::Qwen2VLVideoProcessingTest::test_can_compile_fast_video_processor"
+        ],
+        "multi-gpu": [
+            "tests/models/qwen2_vl/test_modeling_qwen2_vl.py::Qwen2VLModelTest::test_generate_continue_from_past_key_values",
+            "tests/models/qwen2_vl/test_video_processing_qwen2_vl.py::Qwen2VLVideoProcessingTest::test_can_compile_fast_video_processor"
+        ]
+    },
+    "qwen3": {
+        "single-gpu": [
+            "tests/models/qwen3/test_modeling_qwen3.py::Qwen3IntegrationTest::test_export_static_cache"
+        ],
+        "multi-gpu": [
+            "tests/models/qwen3/test_modeling_qwen3.py::Qwen3IntegrationTest::test_export_static_cache"
+        ]
+    },
+    "sam": {
+        "single-gpu": [
+            "tests/models/sam/test_modeling_sam.py::SamVisionModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/sam/test_modeling_sam.py::SamVisionModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "sam_hq": {
+        "single-gpu": [
+            "tests/models/sam_hq/test_modeling_sam_hq.py::SamHQVisionModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/sam_hq/test_modeling_sam_hq.py::SamHQVisionModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "seamless_m4t_v2": {
+        "single-gpu": [
+            "tests/models/seamless_m4t_v2/test_modeling_seamless_m4t_v2.py::SeamlessM4Tv2ModelWithTextInputTest::test_retain_grad_hidden_states_attentions",
+            "tests/models/seamless_m4t_v2/test_modeling_seamless_m4t_v2.py::SeamlessM4Tv2ModelIntegrationTest::test_speech_to_speech_model",
+            "tests/models/seamless_m4t_v2/test_modeling_seamless_m4t_v2.py::SeamlessM4Tv2ModelIntegrationTest::test_text_to_speech_model"
+        ],
+        "multi-gpu": [
+            "tests/models/seamless_m4t_v2/test_modeling_seamless_m4t_v2.py::SeamlessM4Tv2ModelIntegrationTest::test_speech_to_speech_model",
+            "tests/models/seamless_m4t_v2/test_modeling_seamless_m4t_v2.py::SeamlessM4Tv2ModelIntegrationTest::test_text_to_speech_model"
+        ]
+    },
+    "speecht5": {
+        "single-gpu": [
+            "tests/models/speecht5/test_modeling_speecht5.py::SpeechT5ForTextToSpeechIntegrationTests::test_batch_generation",
+            "tests/models/speecht5/test_modeling_speecht5.py::SpeechT5ForTextToSpeechIntegrationTests::test_one_to_many_generation"
+        ],
+        "multi-gpu": [
+            "tests/models/speecht5/test_modeling_speecht5.py::SpeechT5ForTextToSpeechIntegrationTests::test_batch_generation",
+            "tests/models/speecht5/test_modeling_speecht5.py::SpeechT5ForTextToSpeechIntegrationTests::test_one_to_many_generation"
+        ]
+    },
+    "superpoint": {
+        "single-gpu": [],
+        "multi-gpu": [
+            "tests/models/superpoint/test_modeling_superpoint.py::SuperPointModelTest::test_multi_gpu_data_parallel_forward"
+        ]
+    },
+    "tapas": {
+        "single-gpu": [
+            "tests/models/tapas/test_modeling_tapas.py::TapasModelIntegrationTest::test_training_question_answering_head_weak_supervision"
+        ],
+        "multi-gpu": [
+            "tests/models/tapas/test_modeling_tapas.py::TapasModelIntegrationTest::test_training_question_answering_head_weak_supervision"
+        ]
+    },
+    "vitpose": {
+        "single-gpu": [
+            "tests/models/vitpose/test_modeling_vitpose.py::VitPoseModelIntegrationTest::test_batched_inference",
+            "tests/models/vitpose/test_modeling_vitpose.py::VitPoseModelIntegrationTest::test_inference_pose_estimation"
+        ],
+        "multi-gpu": [
+            "tests/models/vitpose/test_modeling_vitpose.py::VitPoseModelIntegrationTest::test_batched_inference",
+            "tests/models/vitpose/test_modeling_vitpose.py::VitPoseModelIntegrationTest::test_inference_pose_estimation"
+        ]
+    },
+    "zamba2": {
+        "single-gpu": [
+            "tests/models/zamba2/test_modeling_zamba2.py::Zamba2ModelTest::test_sdpa_can_dispatch_on_flash"
+        ],
+        "multi-gpu": [
+            "tests/models/zamba2/test_modeling_zamba2.py::Zamba2ModelTest::test_sdpa_can_dispatch_on_flash"
+        ]
+    },
+    "pipelines": {
+        "single-gpu": [
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_simple_s2t",
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_small_model_pt",
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_small_model_pt_bf16",
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_small_model_pt_fp16",
+            "tests/pipelines/test_pipelines_image_segmentation.py::ImageSegmentationPipelineTests::test_integration_torch_image_segmentation",
+            "tests/pipelines/test_pipelines_image_segmentation.py::ImageSegmentationPipelineTests::test_threshold"
+        ],
+        "multi-gpu": [
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_simple_s2t",
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_small_model_pt",
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_small_model_pt_bf16",
+            "tests/pipelines/test_pipelines_automatic_speech_recognition.py::AutomaticSpeechRecognitionPipelineTests::test_small_model_pt_fp16",
+            "tests/pipelines/test_pipelines_image_segmentation.py::ImageSegmentationPipelineTests::test_integration_torch_image_segmentation",
+            "tests/pipelines/test_pipelines_image_segmentation.py::ImageSegmentationPipelineTests::test_threshold"
+        ]
+    },
+    "quantization": {
+        "single-gpu": [
+            "tests/quantization/bitnet_integration/test_bitnet.py::BitNetTest::test_quantized_model",
+            "tests/quantization/finegrained_fp8/test_fp8.py::FP8LinearTest::test_linear_preserves_shape",
+            "tests/quantization/finegrained_fp8/test_fp8.py::FP8LinearTest::test_linear_with_diff_feature_size_preserves_shape"
+        ],
+        "multi-gpu": [
+            "tests/quantization/bitnet_integration/test_bitnet.py::BitNetTest::test_quantized_model",
+            "tests/quantization/finegrained_fp8/test_fp8.py::FP8LinearTest::test_linear_preserves_shape",
+            "tests/quantization/finegrained_fp8/test_fp8.py::FP8LinearTest::test_linear_with_diff_feature_size_preserves_shape"
+        ]
+    },
+    "tensor_parallel": {
+        "single-gpu": [
+            "tests/tensor_parallel/test_tensor_parallel.py::TestTensorParallel::test_model_save"
+        ],
+        "multi-gpu": []
+    },
+    "trainer": {
+        "single-gpu": [
+            "tests/trainer/test_trainer.py::TrainerHyperParameterOptunaIntegrationTest::test_hyperparameter_search",
+            "tests/trainer/test_trainer.py::TrainerHyperParameterMultiObjectOptunaIntegrationTest::test_hyperparameter_search",
+            "tests/trainer/test_trainer.py::TrainerHyperParameterOptunaIntegrationTestWithFullEval::test_hyperparameter_search"
+        ],
+        "multi-gpu": [
+            "tests/trainer/test_trainer.py::TrainerHyperParameterOptunaIntegrationTest::test_hyperparameter_search",
+            "tests/trainer/test_trainer.py::TrainerHyperParameterMultiObjectOptunaIntegrationTest::test_hyperparameter_search"
+        ]
+    }
+}