hf-transformers-bot commited on
Commit
1d4ef54
·
verified ·
1 Parent(s): 7575704

Upload 2025-11-04/ci_results_run_models_gpu/new_failures.json with huggingface_hub

Browse files
2025-11-04/ci_results_run_models_gpu/new_failures.json ADDED
@@ -0,0 +1,707 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bamba": {
3
+ "single-gpu": [
4
+ "tests/models/bamba/test_modeling_bamba.py::BambaModelTest::test_flash_attn_2_inference_equivalence",
5
+ "tests/models/bamba/test_modeling_bamba.py::BambaModelTest::test_flash_attn_2_inference_equivalence_right_padding"
6
+ ],
7
+ "multi-gpu": [
8
+ "tests/models/bamba/test_modeling_bamba.py::BambaModelTest::test_flash_attn_2_inference_equivalence",
9
+ "tests/models/bamba/test_modeling_bamba.py::BambaModelTest::test_flash_attn_2_inference_equivalence_right_padding"
10
+ ]
11
+ },
12
+ "bark": {
13
+ "single-gpu": [
14
+ "tests/models/bark/test_modeling_bark.py::BarkSemanticModelTest::test_eager_matches_fa2_generate",
15
+ "tests/models/bark/test_modeling_bark.py::BarkSemanticModelTest::test_flash_attention_2_continue_generate_with_position_ids",
16
+ "tests/models/bark/test_modeling_bark.py::BarkCoarseModelTest::test_eager_matches_fa2_generate",
17
+ "tests/models/bark/test_modeling_bark.py::BarkCoarseModelTest::test_flash_attention_2_continue_generate_with_position_ids"
18
+ ],
19
+ "multi-gpu": [
20
+ "tests/models/bark/test_modeling_bark.py::BarkSemanticModelTest::test_eager_matches_fa2_generate",
21
+ "tests/models/bark/test_modeling_bark.py::BarkSemanticModelTest::test_flash_attention_2_continue_generate_with_position_ids",
22
+ "tests/models/bark/test_modeling_bark.py::BarkCoarseModelTest::test_eager_matches_fa2_generate",
23
+ "tests/models/bark/test_modeling_bark.py::BarkCoarseModelTest::test_flash_attention_2_continue_generate_with_position_ids"
24
+ ]
25
+ },
26
+ "bert": {
27
+ "single-gpu": [],
28
+ "multi-gpu": [
29
+ "tests/models/bert/test_modeling_bert.py::BertModelTest::test_flash_attn_2_inference_equivalence_right_padding"
30
+ ]
31
+ },
32
+ "blt": {
33
+ "single-gpu": [
34
+ "tests/models/blt/test_modeling_blt.py::BltModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
35
+ ],
36
+ "multi-gpu": [
37
+ "tests/models/blt/test_modeling_blt.py::BltModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
38
+ ]
39
+ },
40
+ "chameleon": {
41
+ "single-gpu": [
42
+ "tests/models/chameleon/test_modeling_chameleon.py::ChameleonModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
43
+ "tests/models/chameleon/test_modeling_chameleon.py::ChameleonVision2SeqModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs"
44
+ ],
45
+ "multi-gpu": [
46
+ "tests/models/chameleon/test_modeling_chameleon.py::ChameleonModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
47
+ ]
48
+ },
49
+ "clap": {
50
+ "single-gpu": [
51
+ "tests/models/clap/test_modeling_clap.py::ClapAudioModelTest::test_flash_attn_2_inference_equivalence",
52
+ "tests/models/clap/test_modeling_clap.py::ClapAudioModelTest::test_flash_attn_2_inference_equivalence_right_padding",
53
+ "tests/models/clap/test_modeling_clap.py::ClapModelTest::test_flash_attn_2_inference_equivalence",
54
+ "tests/models/clap/test_modeling_clap.py::ClapModelTest::test_flash_attn_2_inference_equivalence_right_padding"
55
+ ],
56
+ "multi-gpu": [
57
+ "tests/models/clap/test_modeling_clap.py::ClapAudioModelTest::test_flash_attn_2_inference_equivalence",
58
+ "tests/models/clap/test_modeling_clap.py::ClapAudioModelTest::test_flash_attn_2_inference_equivalence_right_padding",
59
+ "tests/models/clap/test_modeling_clap.py::ClapModelTest::test_flash_attn_2_inference_equivalence",
60
+ "tests/models/clap/test_modeling_clap.py::ClapModelTest::test_flash_attn_2_inference_equivalence_right_padding"
61
+ ]
62
+ },
63
+ "colpali": {
64
+ "single-gpu": [
65
+ "tests/models/colpali/test_modeling_colpali.py::ColPaliForRetrievalModelTest::test_flash_attn_2_inference_equivalence",
66
+ "tests/models/colpali/test_modeling_colpali.py::ColPaliForRetrievalModelTest::test_flash_attn_2_inference_equivalence_right_padding"
67
+ ],
68
+ "multi-gpu": [
69
+ "tests/models/colpali/test_modeling_colpali.py::ColPaliForRetrievalModelTest::test_flash_attn_2_inference_equivalence",
70
+ "tests/models/colpali/test_modeling_colpali.py::ColPaliForRetrievalModelTest::test_flash_attn_2_inference_equivalence_right_padding"
71
+ ]
72
+ },
73
+ "data2vec": {
74
+ "single-gpu": [],
75
+ "multi-gpu": [
76
+ "tests/models/data2vec/test_modeling_data2vec_text.py::Data2VecTextModelTest::test_flash_attn_2_inference_equivalence"
77
+ ]
78
+ },
79
+ "deepseek_v2": {
80
+ "single-gpu": [
81
+ "tests/models/deepseek_v2/test_modeling_deepseek_v2.py::DeepseekV2ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
82
+ "tests/models/deepseek_v2/test_modeling_deepseek_v2.py::DeepseekV2ModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break",
83
+ "tests/models/deepseek_v2/test_modeling_deepseek_v2.py::DeepseekV2ModelTest::test_flash_attn_2_fp32_ln"
84
+ ],
85
+ "multi-gpu": [
86
+ "tests/models/deepseek_v2/test_modeling_deepseek_v2.py::DeepseekV2ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
87
+ "tests/models/deepseek_v2/test_modeling_deepseek_v2.py::DeepseekV2ModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break",
88
+ "tests/models/deepseek_v2/test_modeling_deepseek_v2.py::DeepseekV2ModelTest::test_flash_attn_2_fp32_ln"
89
+ ]
90
+ },
91
+ "deepseek_vl_hybrid": {
92
+ "single-gpu": [
93
+ "tests/models/deepseek_vl_hybrid/test_modeling_deepseek_vl_hybrid.py::DeepseekVLHybridModelTest::test_flash_attention_2_continue_generate_with_position_ids",
94
+ "tests/models/deepseek_vl_hybrid/test_modeling_deepseek_vl_hybrid.py::DeepseekVLHybridModelTest::test_flash_attn_2_fp32_ln",
95
+ "tests/models/deepseek_vl_hybrid/test_modeling_deepseek_vl_hybrid.py::DeepseekVLHybridModelTest::test_flash_attn_2_from_config"
96
+ ],
97
+ "multi-gpu": [
98
+ "tests/models/deepseek_vl_hybrid/test_modeling_deepseek_vl_hybrid.py::DeepseekVLHybridModelTest::test_flash_attention_2_continue_generate_with_position_ids",
99
+ "tests/models/deepseek_vl_hybrid/test_modeling_deepseek_vl_hybrid.py::DeepseekVLHybridModelTest::test_flash_attn_2_fp32_ln",
100
+ "tests/models/deepseek_vl_hybrid/test_modeling_deepseek_vl_hybrid.py::DeepseekVLHybridModelTest::test_flash_attn_2_from_config"
101
+ ]
102
+ },
103
+ "diffllama": {
104
+ "single-gpu": [
105
+ "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_flash_attn_2_generate_padding_right",
106
+ "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_flash_attn_2_inference_equivalence",
107
+ "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_flash_attn_2_inference_equivalence_right_padding"
108
+ ],
109
+ "multi-gpu": [
110
+ "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_flash_attn_2_generate_padding_right",
111
+ "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_flash_attn_2_inference_equivalence",
112
+ "tests/models/diffllama/test_modeling_diffllama.py::DiffLlamaModelTest::test_flash_attn_2_inference_equivalence_right_padding"
113
+ ]
114
+ },
115
+ "donut": {
116
+ "single-gpu": [
117
+ "tests/models/donut/test_modeling_donut_swin.py::DonutSwinModelTest::test_flash_attn_2_inference_equivalence",
118
+ "tests/models/donut/test_modeling_donut_swin.py::DonutSwinModelTest::test_flash_attn_2_inference_equivalence_right_padding"
119
+ ],
120
+ "multi-gpu": [
121
+ "tests/models/donut/test_modeling_donut_swin.py::DonutSwinModelTest::test_flash_attn_2_inference_equivalence",
122
+ "tests/models/donut/test_modeling_donut_swin.py::DonutSwinModelTest::test_flash_attn_2_inference_equivalence_right_padding"
123
+ ]
124
+ },
125
+ "efficientloftr": {
126
+ "single-gpu": [
127
+ "tests/models/efficientloftr/test_modeling_efficientloftr.py::EfficientLoFTRModelTest::test_flash_attn_2_inference_equivalence",
128
+ "tests/models/efficientloftr/test_modeling_efficientloftr.py::EfficientLoFTRModelTest::test_flash_attn_2_inference_equivalence_right_padding"
129
+ ],
130
+ "multi-gpu": [
131
+ "tests/models/efficientloftr/test_modeling_efficientloftr.py::EfficientLoFTRModelTest::test_flash_attn_2_inference_equivalence",
132
+ "tests/models/efficientloftr/test_modeling_efficientloftr.py::EfficientLoFTRModelTest::test_flash_attn_2_inference_equivalence_right_padding"
133
+ ]
134
+ },
135
+ "electra": {
136
+ "single-gpu": [
137
+ "tests/models/electra/test_modeling_electra.py::ElectraModelTest::test_flash_attn_2_inference_equivalence_right_padding"
138
+ ],
139
+ "multi-gpu": []
140
+ },
141
+ "emu3": {
142
+ "single-gpu": [
143
+ "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_flash_attn_2_inference_equivalence",
144
+ "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
145
+ ],
146
+ "multi-gpu": [
147
+ "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_flash_attn_2_inference_equivalence",
148
+ "tests/models/emu3/test_modeling_emu3.py::Emu3Vision2TextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
149
+ ]
150
+ },
151
+ "ernie4_5_moe": {
152
+ "single-gpu": [],
153
+ "multi-gpu": [
154
+ "tests/models/ernie4_5_moe/test_modeling_ernie4_5_moe.py::Ernie4_5_MoeModelTest::test_flash_attn_2_equivalence"
155
+ ]
156
+ },
157
+ "exaone4": {
158
+ "single-gpu": [
159
+ "tests/models/exaone4/test_modeling_exaone4.py::Exaone4ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
160
+ ],
161
+ "multi-gpu": [
162
+ "tests/models/exaone4/test_modeling_exaone4.py::Exaone4ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
163
+ "tests/models/exaone4/test_modeling_exaone4.py::Exaone4IntegrationTest::test_model_generation_long_flash"
164
+ ]
165
+ },
166
+ "falcon": {
167
+ "single-gpu": [
168
+ "tests/models/falcon/test_modeling_falcon.py::FalconModelTest::test_flash_attn_2_inference_equivalence",
169
+ "tests/models/falcon/test_modeling_falcon.py::FalconModelTest::test_flash_attn_2_inference_equivalence_right_padding"
170
+ ],
171
+ "multi-gpu": [
172
+ "tests/models/falcon/test_modeling_falcon.py::FalconModelTest::test_flash_attn_2_inference_equivalence",
173
+ "tests/models/falcon/test_modeling_falcon.py::FalconModelTest::test_flash_attn_2_inference_equivalence_right_padding"
174
+ ]
175
+ },
176
+ "flex_olmo": {
177
+ "single-gpu": [
178
+ "tests/models/flex_olmo/test_modeling_flex_olmo.py::FlexOlmoModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
179
+ "tests/models/flex_olmo/test_modeling_flex_olmo.py::FlexOlmoModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
180
+ ],
181
+ "multi-gpu": [
182
+ "tests/models/flex_olmo/test_modeling_flex_olmo.py::FlexOlmoModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
183
+ "tests/models/flex_olmo/test_modeling_flex_olmo.py::FlexOlmoModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break",
184
+ "tests/models/flex_olmo/test_modeling_flex_olmo.py::FlexOlmoModelTest::test_flash_attn_2_equivalence"
185
+ ]
186
+ },
187
+ "gemma3n": {
188
+ "single-gpu": [
189
+ "tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_flash_attn_2_equivalence",
190
+ "tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_flash_attn_2_inference_equivalence",
191
+ "tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
192
+ ],
193
+ "multi-gpu": [
194
+ "tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_flash_attn_2_equivalence",
195
+ "tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_flash_attn_2_inference_equivalence",
196
+ "tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
197
+ ]
198
+ },
199
+ "glm4_moe": {
200
+ "single-gpu": [
201
+ "tests/models/glm4_moe/test_modeling_glm4_moe.py::Glm4MoeModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
202
+ ],
203
+ "multi-gpu": [
204
+ "tests/models/glm4_moe/test_modeling_glm4_moe.py::Glm4MoeModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
205
+ ]
206
+ },
207
+ "gpt2": {
208
+ "single-gpu": [
209
+ "tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelLanguageGenerationTest::test_flash_attn_2_generate_padding_left"
210
+ ],
211
+ "multi-gpu": [
212
+ "tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelLanguageGenerationTest::test_flash_attn_2_generate_padding_left"
213
+ ]
214
+ },
215
+ "gpt_oss": {
216
+ "single-gpu": [
217
+ "tests/models/gpt_oss/test_modeling_gpt_oss.py::GptOssModelTest::test_flash_attn_2_inference_equivalence",
218
+ "tests/models/gpt_oss/test_modeling_gpt_oss.py::GptOssModelTest::test_flash_attn_2_inference_equivalence_right_padding"
219
+ ],
220
+ "multi-gpu": [
221
+ "tests/models/gpt_oss/test_modeling_gpt_oss.py::GptOssModelTest::test_flash_attn_2_inference_equivalence",
222
+ "tests/models/gpt_oss/test_modeling_gpt_oss.py::GptOssModelTest::test_flash_attn_2_inference_equivalence_right_padding"
223
+ ]
224
+ },
225
+ "granitemoe": {
226
+ "single-gpu": [
227
+ "tests/models/granitemoe/test_modeling_granitemoe.py::GraniteMoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs"
228
+ ],
229
+ "multi-gpu": [
230
+ "tests/models/granitemoe/test_modeling_granitemoe.py::GraniteMoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs"
231
+ ]
232
+ },
233
+ "granitemoehybrid": {
234
+ "single-gpu": [
235
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::BambaModelTest::test_flash_attn_2_inference_equivalence",
236
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::BambaModelTest::test_flash_attn_2_inference_equivalence_right_padding",
237
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::GraniteMoeHybridModelTest::test_flash_attn_2_inference_equivalence",
238
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::GraniteMoeHybridModelTest::test_flash_attn_2_inference_equivalence_right_padding"
239
+ ],
240
+ "multi-gpu": [
241
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::BambaModelTest::test_flash_attn_2_inference_equivalence",
242
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::BambaModelTest::test_flash_attn_2_inference_equivalence_right_padding",
243
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::GraniteMoeHybridModelTest::test_flash_attn_2_inference_equivalence",
244
+ "tests/models/granitemoehybrid/test_modeling_granitemoehybrid.py::GraniteMoeHybridModelTest::test_flash_attn_2_inference_equivalence_right_padding"
245
+ ]
246
+ },
247
+ "granitemoeshared": {
248
+ "single-gpu": [
249
+ "tests/models/granitemoeshared/test_modeling_granitemoeshared.py::GraniteMoeSharedModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs"
250
+ ],
251
+ "multi-gpu": [
252
+ "tests/models/granitemoeshared/test_modeling_granitemoeshared.py::GraniteMoeSharedModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs"
253
+ ]
254
+ },
255
+ "grounding_dino": {
256
+ "single-gpu": [
257
+ "tests/models/grounding_dino/test_modeling_grounding_dino.py::GroundingDinoModelTest::test_flash_attn_2_inference_equivalence",
258
+ "tests/models/grounding_dino/test_modeling_grounding_dino.py::GroundingDinoModelTest::test_flash_attn_2_inference_equivalence_right_padding"
259
+ ],
260
+ "multi-gpu": [
261
+ "tests/models/grounding_dino/test_modeling_grounding_dino.py::GroundingDinoModelTest::test_flash_attn_2_inference_equivalence",
262
+ "tests/models/grounding_dino/test_modeling_grounding_dino.py::GroundingDinoModelTest::test_flash_attn_2_inference_equivalence_right_padding"
263
+ ]
264
+ },
265
+ "instructblip": {
266
+ "single-gpu": [
267
+ "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_fp32_ln",
268
+ "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_from_config"
269
+ ],
270
+ "multi-gpu": [
271
+ "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_fp32_ln",
272
+ "tests/models/instructblip/test_modeling_instructblip.py::InstructBlipForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_from_config"
273
+ ]
274
+ },
275
+ "instructblipvideo": {
276
+ "single-gpu": [
277
+ "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_fp32_ln",
278
+ "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_from_config"
279
+ ],
280
+ "multi-gpu": [
281
+ "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_fp32_ln",
282
+ "tests/models/instructblipvideo/test_modeling_instructblipvideo.py::InstructBlipVideoForConditionalGenerationDecoderOnlyTest::test_flash_attn_2_from_config"
283
+ ]
284
+ },
285
+ "janus": {
286
+ "single-gpu": [
287
+ "tests/models/janus/test_modeling_janus.py::JanusVisionText2TextModelTest::test_flash_attn_2_inference_equivalence",
288
+ "tests/models/janus/test_modeling_janus.py::JanusVisionText2TextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
289
+ ],
290
+ "multi-gpu": [
291
+ "tests/models/janus/test_modeling_janus.py::JanusVisionText2TextModelTest::test_flash_attn_2_inference_equivalence",
292
+ "tests/models/janus/test_modeling_janus.py::JanusVisionText2TextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
293
+ ]
294
+ },
295
+ "jetmoe": {
296
+ "single-gpu": [
297
+ "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeModelTest::test_flash_attn_2_equivalence",
298
+ "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeModelTest::test_flash_attn_2_fp32_ln"
299
+ ],
300
+ "multi-gpu": [
301
+ "tests/models/jetmoe/test_modeling_jetmoe.py::JetMoeModelTest::test_flash_attn_2_fp32_ln"
302
+ ]
303
+ },
304
+ "kosmos2": {
305
+ "single-gpu": [
306
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_eager_matches_fa2_generate",
307
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attention_2_continue_generate_with_position_ids",
308
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_can_dispatch_composite_models",
309
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_fp32_ln",
310
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_from_config",
311
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_inference_equivalence",
312
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_inference_equivalence_right_padding"
313
+ ],
314
+ "multi-gpu": [
315
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_eager_matches_fa2_generate",
316
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attention_2_continue_generate_with_position_ids",
317
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_can_dispatch_composite_models",
318
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_fp32_ln",
319
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_from_config",
320
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_inference_equivalence",
321
+ "tests/models/kosmos2/test_modeling_kosmos2.py::Kosmos2ModelTest::test_flash_attn_2_inference_equivalence_right_padding"
322
+ ]
323
+ },
324
+ "kosmos2_5": {
325
+ "single-gpu": [
326
+ "tests/models/kosmos2_5/test_modeling_kosmos2_5.py::Kosmos2_5ModelTest::test_flash_attn_2_can_dispatch_composite_models"
327
+ ],
328
+ "multi-gpu": [
329
+ "tests/models/kosmos2_5/test_modeling_kosmos2_5.py::Kosmos2_5ModelTest::test_flash_attn_2_can_dispatch_composite_models"
330
+ ]
331
+ },
332
+ "kyutai_speech_to_text": {
333
+ "single-gpu": [
334
+ "tests/models/kyutai_speech_to_text/test_modeling_kyutai_speech_to_text.py::KyutaiSpeechToTextModelTest::test_eager_matches_fa2_generate",
335
+ "tests/models/kyutai_speech_to_text/test_modeling_kyutai_speech_to_text.py::KyutaiSpeechToTextModelTest::test_flash_attn_2_inference_equivalence",
336
+ "tests/models/kyutai_speech_to_text/test_modeling_kyutai_speech_to_text.py::KyutaiSpeechToTextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
337
+ ],
338
+ "multi-gpu": [
339
+ "tests/models/kyutai_speech_to_text/test_modeling_kyutai_speech_to_text.py::KyutaiSpeechToTextModelTest::test_eager_matches_fa2_generate",
340
+ "tests/models/kyutai_speech_to_text/test_modeling_kyutai_speech_to_text.py::KyutaiSpeechToTextModelTest::test_flash_attn_2_inference_equivalence",
341
+ "tests/models/kyutai_speech_to_text/test_modeling_kyutai_speech_to_text.py::KyutaiSpeechToTextModelTest::test_flash_attn_2_inference_equivalence_right_padding"
342
+ ]
343
+ },
344
+ "lfm2": {
345
+ "single-gpu": [
346
+ "tests/models/lfm2/test_modeling_lfm2.py::Lfm2ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
347
+ "tests/models/lfm2/test_modeling_lfm2.py::Lfm2ModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
348
+ ],
349
+ "multi-gpu": [
350
+ "tests/models/lfm2/test_modeling_lfm2.py::Lfm2ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
351
+ "tests/models/lfm2/test_modeling_lfm2.py::Lfm2ModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
352
+ ]
353
+ },
354
+ "lfm2_moe": {
355
+ "single-gpu": [
356
+ "tests/models/lfm2_moe/test_modeling_lfm2_moe.py::Lfm2MoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
357
+ "tests/models/lfm2_moe/test_modeling_lfm2_moe.py::Lfm2MoeModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
358
+ ],
359
+ "multi-gpu": [
360
+ "tests/models/lfm2_moe/test_modeling_lfm2_moe.py::Lfm2MoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
361
+ "tests/models/lfm2_moe/test_modeling_lfm2_moe.py::Lfm2MoeModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
362
+ ]
363
+ },
364
+ "lfm2_vl": {
365
+ "single-gpu": [
366
+ "tests/models/lfm2_vl/test_modeling_lfm2_vl.py::Lfm2VlModelTest::test_flash_attn_2_inference_equivalence",
367
+ "tests/models/lfm2_vl/test_modeling_lfm2_vl.py::Lfm2VlModelTest::test_flash_attn_2_inference_equivalence_right_padding"
368
+ ],
369
+ "multi-gpu": [
370
+ "tests/models/lfm2_vl/test_modeling_lfm2_vl.py::Lfm2VlModelTest::test_flash_attn_2_inference_equivalence",
371
+ "tests/models/lfm2_vl/test_modeling_lfm2_vl.py::Lfm2VlModelTest::test_flash_attn_2_inference_equivalence_right_padding"
372
+ ]
373
+ },
374
+ "llava_next": {
375
+ "single-gpu": [
376
+ "tests/models/llava_next/test_modeling_llava_next.py::LlavaNextForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
377
+ "tests/models/llava_next/test_modeling_llava_next.py::LlavaNextForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
378
+ ],
379
+ "multi-gpu": [
380
+ "tests/models/llava_next/test_modeling_llava_next.py::LlavaNextForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
381
+ "tests/models/llava_next/test_modeling_llava_next.py::LlavaNextForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
382
+ ]
383
+ },
384
+ "llava_next_video": {
385
+ "single-gpu": [
386
+ "tests/models/llava_next_video/test_modeling_llava_next_video.py::LlavaNextVideoForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
387
+ "tests/models/llava_next_video/test_modeling_llava_next_video.py::LlavaNextVideoForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
388
+ ],
389
+ "multi-gpu": [
390
+ "tests/models/llava_next_video/test_modeling_llava_next_video.py::LlavaNextVideoForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
391
+ "tests/models/llava_next_video/test_modeling_llava_next_video.py::LlavaNextVideoForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
392
+ ]
393
+ },
394
+ "llava_onevision": {
395
+ "single-gpu": [
396
+ "tests/models/llava_onevision/test_modeling_llava_onevision.py::LlavaOnevisionForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
397
+ "tests/models/llava_onevision/test_modeling_llava_onevision.py::LlavaOnevisionForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
398
+ ],
399
+ "multi-gpu": [
400
+ "tests/models/llava_onevision/test_modeling_llava_onevision.py::LlavaOnevisionForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
401
+ "tests/models/llava_onevision/test_modeling_llava_onevision.py::LlavaOnevisionForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
402
+ ]
403
+ },
404
+ "mask2former": {
405
+ "single-gpu": [
406
+ "tests/models/mask2former/test_modeling_mask2former.py::Mask2FormerModelTest::test_flash_attn_2_inference_equivalence",
407
+ "tests/models/mask2former/test_modeling_mask2former.py::Mask2FormerModelTest::test_flash_attn_2_inference_equivalence_right_padding"
408
+ ],
409
+ "multi-gpu": [
410
+ "tests/models/mask2former/test_modeling_mask2former.py::Mask2FormerModelTest::test_flash_attn_2_inference_equivalence",
411
+ "tests/models/mask2former/test_modeling_mask2former.py::Mask2FormerModelTest::test_flash_attn_2_inference_equivalence_right_padding"
412
+ ]
413
+ },
414
+ "maskformer": {
415
+ "single-gpu": [
416
+ "tests/models/maskformer/test_modeling_maskformer.py::MaskFormerModelTest::test_flash_attn_2_inference_equivalence",
417
+ "tests/models/maskformer/test_modeling_maskformer.py::MaskFormerModelTest::test_flash_attn_2_inference_equivalence_right_padding",
418
+ "tests/models/maskformer/test_modeling_maskformer_swin.py::MaskFormerSwinModelTest::test_flash_attn_2_inference_equivalence",
419
+ "tests/models/maskformer/test_modeling_maskformer_swin.py::MaskFormerSwinModelTest::test_flash_attn_2_inference_equivalence_right_padding"
420
+ ],
421
+ "multi-gpu": [
422
+ "tests/models/maskformer/test_modeling_maskformer.py::MaskFormerModelTest::test_flash_attn_2_inference_equivalence",
423
+ "tests/models/maskformer/test_modeling_maskformer.py::MaskFormerModelTest::test_flash_attn_2_inference_equivalence_right_padding",
424
+ "tests/models/maskformer/test_modeling_maskformer_swin.py::MaskFormerSwinModelTest::test_flash_attn_2_inference_equivalence",
425
+ "tests/models/maskformer/test_modeling_maskformer_swin.py::MaskFormerSwinModelTest::test_flash_attn_2_inference_equivalence_right_padding"
426
+ ]
427
+ },
428
+ "mixtral": {
429
+ "single-gpu": [
430
+ "tests/models/mixtral/test_modeling_mixtral.py::MixtralModelTest::test_flash_attn_2_equivalence"
431
+ ],
432
+ "multi-gpu": [
433
+ "tests/models/mixtral/test_modeling_mixtral.py::MixtralModelTest::test_flash_attn_2_equivalence"
434
+ ]
435
+ },
436
+ "mllama": {
437
+ "single-gpu": [
438
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForCausalLMModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
439
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_eager_matches_fa2_generate",
440
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
441
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
442
+ ],
443
+ "multi-gpu": [
444
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_eager_matches_fa2_generate",
445
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
446
+ "tests/models/mllama/test_modeling_mllama.py::MllamaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
447
+ ]
448
+ },
449
+ "mm_grounding_dino": {
450
+ "single-gpu": [
451
+ "tests/models/mm_grounding_dino/test_modeling_mm_grounding_dino.py::MMGroundingDinoModelTest::test_flash_attn_2_inference_equivalence",
452
+ "tests/models/mm_grounding_dino/test_modeling_mm_grounding_dino.py::MMGroundingDinoModelTest::test_flash_attn_2_inference_equivalence_right_padding"
453
+ ],
454
+ "multi-gpu": [
455
+ "tests/models/mm_grounding_dino/test_modeling_mm_grounding_dino.py::MMGroundingDinoModelTest::test_flash_attn_2_inference_equivalence",
456
+ "tests/models/mm_grounding_dino/test_modeling_mm_grounding_dino.py::MMGroundingDinoModelTest::test_flash_attn_2_inference_equivalence_right_padding"
457
+ ]
458
+ },
459
+ "moshi": {
460
+ "single-gpu": [
461
+ "tests/models/moshi/test_modeling_moshi.py::MoshiDecoderTest::test_flash_attn_2_inference_equivalence",
462
+ "tests/models/moshi/test_modeling_moshi.py::MoshiDecoderTest::test_flash_attn_2_inference_equivalence_right_padding",
463
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_eager_matches_fa2_generate",
464
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_fp32_ln",
465
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_from_config",
466
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_inference_equivalence",
467
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_inference_equivalence_right_padding"
468
+ ],
469
+ "multi-gpu": [
470
+ "tests/models/moshi/test_modeling_moshi.py::MoshiDecoderTest::test_flash_attn_2_inference_equivalence",
471
+ "tests/models/moshi/test_modeling_moshi.py::MoshiDecoderTest::test_flash_attn_2_inference_equivalence_right_padding",
472
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_eager_matches_fa2_generate",
473
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_fp32_ln",
474
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_from_config",
475
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_inference_equivalence",
476
+ "tests/models/moshi/test_modeling_moshi.py::MoshiTest::test_flash_attn_2_inference_equivalence_right_padding"
477
+ ]
478
+ },
479
+ "nemotron": {
480
+ "single-gpu": [
481
+ "tests/models/nemotron/test_modeling_nemotron.py::NemotronModelTest::test_flash_attn_2_equivalence"
482
+ ],
483
+ "multi-gpu": [
484
+ "tests/models/nemotron/test_modeling_nemotron.py::NemotronModelTest::test_flash_attn_2_equivalence"
485
+ ]
486
+ },
487
+ "olmo": {
488
+ "single-gpu": [
489
+ "tests/models/olmo/test_modeling_olmo.py::OlmoModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
490
+ ],
491
+ "multi-gpu": [
492
+ "tests/models/olmo/test_modeling_olmo.py::OlmoModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
493
+ ]
494
+ },
495
+ "olmo2": {
496
+ "single-gpu": [
497
+ "tests/models/olmo2/test_modeling_olmo2.py::Olmo2ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
498
+ ],
499
+ "multi-gpu": [
500
+ "tests/models/olmo2/test_modeling_olmo2.py::Olmo2ModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
501
+ ]
502
+ },
503
+ "omdet_turbo": {
504
+ "single-gpu": [
505
+ "tests/models/omdet_turbo/test_modeling_omdet_turbo.py::OmDetTurboModelTest::test_flash_attn_2_inference_equivalence",
506
+ "tests/models/omdet_turbo/test_modeling_omdet_turbo.py::OmDetTurboModelTest::test_flash_attn_2_inference_equivalence_right_padding"
507
+ ],
508
+ "multi-gpu": [
509
+ "tests/models/omdet_turbo/test_modeling_omdet_turbo.py::OmDetTurboModelTest::test_flash_attn_2_inference_equivalence",
510
+ "tests/models/omdet_turbo/test_modeling_omdet_turbo.py::OmDetTurboModelTest::test_flash_attn_2_inference_equivalence_right_padding"
511
+ ]
512
+ },
513
+ "oneformer": {
514
+ "single-gpu": [
515
+ "tests/models/oneformer/test_modeling_oneformer.py::OneFormerModelTest::test_flash_attn_2_inference_equivalence",
516
+ "tests/models/oneformer/test_modeling_oneformer.py::OneFormerModelTest::test_flash_attn_2_inference_equivalence_right_padding"
517
+ ],
518
+ "multi-gpu": [
519
+ "tests/models/oneformer/test_modeling_oneformer.py::OneFormerModelTest::test_flash_attn_2_inference_equivalence",
520
+ "tests/models/oneformer/test_modeling_oneformer.py::OneFormerModelTest::test_flash_attn_2_inference_equivalence_right_padding"
521
+ ]
522
+ },
523
+ "paligemma": {
524
+ "single-gpu": [
525
+ "tests/models/paligemma/test_modeling_paligemma.py::PaliGemmaForConditionalGenerationModelTest::test_flash_attn_2_from_config",
526
+ "tests/models/paligemma/test_modeling_paligemma.py::PaliGemmaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
527
+ "tests/models/paligemma/test_modeling_paligemma.py::PaliGemmaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
528
+ ],
529
+ "multi-gpu": [
530
+ "tests/models/paligemma/test_modeling_paligemma.py::PaliGemmaForConditionalGenerationModelTest::test_flash_attn_2_from_config",
531
+ "tests/models/paligemma/test_modeling_paligemma.py::PaliGemmaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
532
+ "tests/models/paligemma/test_modeling_paligemma.py::PaliGemmaForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
533
+ ]
534
+ },
535
+ "paligemma2": {
536
+ "single-gpu": [
537
+ "tests/models/paligemma2/test_modeling_paligemma2.py::PaliGemma2ForConditionalGenerationModelTest::test_flash_attn_2_from_config",
538
+ "tests/models/paligemma2/test_modeling_paligemma2.py::PaliGemma2ForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
539
+ "tests/models/paligemma2/test_modeling_paligemma2.py::PaliGemma2ForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
540
+ ],
541
+ "multi-gpu": [
542
+ "tests/models/paligemma2/test_modeling_paligemma2.py::PaliGemma2ForConditionalGenerationModelTest::test_flash_attn_2_from_config",
543
+ "tests/models/paligemma2/test_modeling_paligemma2.py::PaliGemma2ForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence",
544
+ "tests/models/paligemma2/test_modeling_paligemma2.py::PaliGemma2ForConditionalGenerationModelTest::test_flash_attn_2_inference_equivalence_right_padding"
545
+ ]
546
+ },
547
+ "pegasus_x": {
548
+ "single-gpu": [
549
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXModelTest::test_flash_attn_2_inference_equivalence",
550
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXModelTest::test_flash_attn_2_inference_equivalence_right_padding",
551
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXStandaloneDecoderModelTest::test_flash_attn_2_inference_equivalence",
552
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXStandaloneDecoderModelTest::test_flash_attn_2_inference_equivalence_right_padding"
553
+ ],
554
+ "multi-gpu": [
555
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXModelTest::test_flash_attn_2_inference_equivalence",
556
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXModelTest::test_flash_attn_2_inference_equivalence_right_padding",
557
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXStandaloneDecoderModelTest::test_flash_attn_2_inference_equivalence",
558
+ "tests/models/pegasus_x/test_modeling_pegasus_x.py::PegasusXStandaloneDecoderModelTest::test_flash_attn_2_inference_equivalence_right_padding"
559
+ ]
560
+ },
561
+ "perception_lm": {
562
+ "single-gpu": [
563
+ "tests/models/perception_lm/test_modeling_perception_lm.py::PerceptionLMForConditionalGenerationModelTest::test_flash_attention_2_continue_generate_with_position_ids"
564
+ ],
565
+ "multi-gpu": [
566
+ "tests/models/perception_lm/test_modeling_perception_lm.py::PerceptionLMForConditionalGenerationModelTest::test_flash_attention_2_continue_generate_with_position_ids"
567
+ ]
568
+ },
569
+ "phi": {
570
+ "single-gpu": [
571
+ "tests/models/phi/test_modeling_phi.py::PhiModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
572
+ ],
573
+ "multi-gpu": [
574
+ "tests/models/phi/test_modeling_phi.py::PhiModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids"
575
+ ]
576
+ },
577
+ "phimoe": {
578
+ "single-gpu": [
579
+ "tests/models/phimoe/test_modeling_phimoe.py::PhimoeModelTest::test_flash_attn_2_equivalence"
580
+ ],
581
+ "multi-gpu": [
582
+ "tests/models/phimoe/test_modeling_phimoe.py::PhimoeModelTest::test_flash_attn_2_equivalence"
583
+ ]
584
+ },
585
+ "pixtral": {
586
+ "single-gpu": [
587
+ "tests/models/pixtral/test_modeling_pixtral.py::PixtralVisionModelModelTest::test_flash_attn_2_inference_equivalence",
588
+ "tests/models/pixtral/test_modeling_pixtral.py::PixtralVisionModelModelTest::test_flash_attn_2_inference_equivalence_right_padding"
589
+ ],
590
+ "multi-gpu": [
591
+ "tests/models/pixtral/test_modeling_pixtral.py::PixtralVisionModelModelTest::test_flash_attn_2_inference_equivalence",
592
+ "tests/models/pixtral/test_modeling_pixtral.py::PixtralVisionModelModelTest::test_flash_attn_2_inference_equivalence_right_padding"
593
+ ]
594
+ },
595
+ "qwen2_5_vl": {
596
+ "single-gpu": [
597
+ "tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test_batch_wo_image_flashatt2"
598
+ ],
599
+ "multi-gpu": [
600
+ "tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test_batch_wo_image_flashatt2"
601
+ ]
602
+ },
603
+ "qwen3_omni_moe": {
604
+ "single-gpu": [
605
+ "tests/models/qwen3_omni_moe/test_modeling_qwen3_omni_moe.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
606
+ "tests/models/qwen3_omni_moe/test_modeling_qwen3_omni_moe.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs",
607
+ "tests/models/qwen3_omni_moe/test_modeling_qwen3_omni_moe.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_batch_flashatt2"
608
+ ],
609
+ "multi-gpu": [
610
+ "tests/models/qwen3_omni_moe/test_modeling_qwen3_omni_moe.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids",
611
+ "tests/models/qwen3_omni_moe/test_modeling_qwen3_omni_moe.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs",
612
+ "tests/models/qwen3_omni_moe/test_modeling_qwen3_omni_moe.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_batch_flashatt2"
613
+ ]
614
+ },
615
+ "roberta_prelayernorm": {
616
+ "single-gpu": [
617
+ "tests/models/roberta_prelayernorm/test_modeling_roberta_prelayernorm.py::RobertaPreLayerNormModelTest::test_flash_attn_2_inference_equivalence_right_padding"
618
+ ],
619
+ "multi-gpu": []
620
+ },
621
+ "sam2": {
622
+ "single-gpu": [
623
+ "tests/models/sam2/test_modeling_sam2.py::Sam2ModelTest::test_flash_attn_2_can_dispatch_composite_models"
624
+ ],
625
+ "multi-gpu": [
626
+ "tests/models/sam2/test_modeling_sam2.py::Sam2ModelTest::test_flash_attn_2_can_dispatch_composite_models"
627
+ ]
628
+ },
629
+ "smollm3": {
630
+ "single-gpu": [
631
+ "tests/models/smollm3/test_modeling_smollm3.py::SmolLM3IntegrationTest::test_model_3b_long_prompt"
632
+ ],
633
+ "multi-gpu": [
634
+ "tests/models/smollm3/test_modeling_smollm3.py::SmolLM3IntegrationTest::test_model_3b_long_prompt"
635
+ ]
636
+ },
637
+ "squeezebert": {
638
+ "single-gpu": [
639
+ "tests/models/squeezebert/test_modeling_squeezebert.py::SqueezeBertModelTest::test_flash_attn_2_inference_equivalence",
640
+ "tests/models/squeezebert/test_modeling_squeezebert.py::SqueezeBertModelTest::test_flash_attn_2_inference_equivalence_right_padding"
641
+ ],
642
+ "multi-gpu": [
643
+ "tests/models/squeezebert/test_modeling_squeezebert.py::SqueezeBertModelTest::test_flash_attn_2_inference_equivalence",
644
+ "tests/models/squeezebert/test_modeling_squeezebert.py::SqueezeBertModelTest::test_flash_attn_2_inference_equivalence_right_padding"
645
+ ]
646
+ },
647
+ "starcoder2": {
648
+ "single-gpu": [
649
+ "tests/models/starcoder2/test_modeling_starcoder2.py::Starcoder2IntegrationTest::test_starcoder2_batched_generation_fa2"
650
+ ],
651
+ "multi-gpu": [
652
+ "tests/models/starcoder2/test_modeling_starcoder2.py::Starcoder2IntegrationTest::test_starcoder2_batched_generation_fa2"
653
+ ]
654
+ },
655
+ "swin": {
656
+ "single-gpu": [
657
+ "tests/models/swin/test_modeling_swin.py::SwinModelTest::test_flash_attn_2_inference_equivalence",
658
+ "tests/models/swin/test_modeling_swin.py::SwinModelTest::test_flash_attn_2_inference_equivalence_right_padding"
659
+ ],
660
+ "multi-gpu": [
661
+ "tests/models/swin/test_modeling_swin.py::SwinModelTest::test_flash_attn_2_inference_equivalence",
662
+ "tests/models/swin/test_modeling_swin.py::SwinModelTest::test_flash_attn_2_inference_equivalence_right_padding"
663
+ ]
664
+ },
665
+ "swin2sr": {
666
+ "single-gpu": [
667
+ "tests/models/swin2sr/test_modeling_swin2sr.py::Swin2SRModelTest::test_flash_attn_2_inference_equivalence",
668
+ "tests/models/swin2sr/test_modeling_swin2sr.py::Swin2SRModelTest::test_flash_attn_2_inference_equivalence_right_padding"
669
+ ],
670
+ "multi-gpu": [
671
+ "tests/models/swin2sr/test_modeling_swin2sr.py::Swin2SRModelTest::test_flash_attn_2_inference_equivalence",
672
+ "tests/models/swin2sr/test_modeling_swin2sr.py::Swin2SRModelTest::test_flash_attn_2_inference_equivalence_right_padding"
673
+ ]
674
+ },
675
+ "swinv2": {
676
+ "single-gpu": [
677
+ "tests/models/swinv2/test_modeling_swinv2.py::Swinv2ModelTest::test_flash_attn_2_inference_equivalence",
678
+ "tests/models/swinv2/test_modeling_swinv2.py::Swinv2ModelTest::test_flash_attn_2_inference_equivalence_right_padding"
679
+ ],
680
+ "multi-gpu": [
681
+ "tests/models/swinv2/test_modeling_swinv2.py::Swinv2ModelTest::test_flash_attn_2_inference_equivalence",
682
+ "tests/models/swinv2/test_modeling_swinv2.py::Swinv2ModelTest::test_flash_attn_2_inference_equivalence_right_padding"
683
+ ]
684
+ },
685
+ "t5gemma": {
686
+ "single-gpu": [
687
+ "tests/models/t5gemma/test_modeling_t5gemma.py::T5GemmaModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
688
+ ],
689
+ "multi-gpu": [
690
+ "tests/models/t5gemma/test_modeling_t5gemma.py::T5GemmaModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break"
691
+ ]
692
+ },
693
+ "xmod": {
694
+ "single-gpu": [],
695
+ "multi-gpu": [
696
+ "tests/models/xmod/test_modeling_xmod.py::XmodModelTest::test_flash_attn_2_inference_equivalence"
697
+ ]
698
+ },
699
+ "zamba": {
700
+ "single-gpu": [
701
+ "tests/models/zamba/test_modeling_zamba.py::ZambaModelTest::test_flash_attn_2_fp32_ln"
702
+ ],
703
+ "multi-gpu": [
704
+ "tests/models/zamba/test_modeling_zamba.py::ZambaModelTest::test_flash_attn_2_fp32_ln"
705
+ ]
706
+ }
707
+ }