-
Notifications
You must be signed in to change notification settings - Fork 166
/
olive_config.json
368 lines (368 loc) · 16.7 KB
/
olive_config.json
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
{
"passes": {
"AppendPrePostProcessingOps": {
"module_path": "olive.passes.onnx.append_pre_post_processing_ops.AppendPrePostProcessingOps",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"DynamicToFixedShape": {
"module_path": "olive.passes.onnx.dynamic_to_fixed_shape.DynamicToFixedShape",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"ExtractAdapters": {
"module_path": "olive.passes.onnx.extract_adapters.ExtractAdapters",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"ModelBuilder": {
"module_path": "olive.passes.onnx.model_builder.ModelBuilder",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "int4", "int8", "fp16", "fp32" ],
"extra_dependencies": [ "ort-genai" ]
},
"IncDynamicQuantization": {
"module_path": "olive.passes.onnx.inc_quantization.IncDynamicQuantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int4", "int8" ],
"extra_dependencies": [ "inc" ]
},
"IncQuantization": {
"module_path": "olive.passes.onnx.inc_quantization.IncQuantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int4", "int8" ],
"extra_dependencies": [ "inc" ]
},
"IncStaticQuantization": {
"module_path": "olive.passes.onnx.inc_quantization.IncStaticQuantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int4", "int8" ],
"extra_dependencies": [ "inc" ]
},
"InsertBeamSearch": {
"module_path": "olive.passes.onnx.insert_beam_search.InsertBeamSearch",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"MatMulNBitsToQDQ": {
"module_path": "olive.passes.onnx.mnb_to_qdq.MatMulNBitsToQDQ",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"MixedPrecisionOverrides": {
"module_path": "olive.passes.onnx.mixed_precision_overrides.MixedPrecisionOverrides",
"supported_providers": [ "QNNExecutionProvider" ],
"supported_accelerators": [ "npu" ],
"supported_precisions": [ "*" ]
},
"MoEExpertsDistributor": {
"module_path": "olive.passes.onnx.moe_experts_distributor.MoEExpertsDistributor",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"OnnxBnb4Quantization": {
"module_path": "olive.passes.onnx.bnb_quantization.OnnxBnb4Quantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "fp4", "nf4" ]
},
"OnnxConversion": {
"module_path": "olive.passes.onnx.conversion.OnnxConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"OnnxDynamicQuantization": {
"module_path": "olive.passes.onnx.quantization.OnnxDynamicQuantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int8", "uint8" ]
},
"OnnxFloatToFloat16": {
"module_path": "olive.passes.onnx.float16_conversion.OnnxFloatToFloat16",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "fp16" ]
},
"OnnxIOFloat16ToFloat32": {
"module_path": "olive.passes.onnx.float32_conversion.OnnxIOFloat16ToFloat32",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "fp32" ]
},
"OnnxMatMul4Quantizer": {
"module_path": "olive.passes.onnx.quantization.OnnxMatMul4Quantizer",
"supported_providers": [ "CPUExecutionProvider", "CUDAExecutionProvider", "DmlExecutionProvider" ],
"supported_accelerators": [ "cpu", "gpu" ],
"supported_precisions": [ "int4" ]
},
"OnnxModelOptimizer": {
"module_path": "olive.passes.onnx.model_optimizer.OnnxModelOptimizer",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"OnnxOpVersionConversion": {
"module_path": "olive.passes.onnx.conversion.OnnxOpVersionConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"OnnxQuantization": {
"module_path": "olive.passes.onnx.quantization.OnnxQuantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int8" ]
},
"OnnxStaticQuantization": {
"module_path": "olive.passes.onnx.quantization.OnnxStaticQuantization",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int8", "int16", "uint8", "uint16" ]
},
"OptimumConversion": {
"module_path": "olive.passes.onnx.optimum_conversion.OptimumConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "optimum" ]
},
"OptimumMerging": {
"module_path": "olive.passes.onnx.optimum_merging.OptimumMerging",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "optimum" ]
},
"OrtMixedPrecision": {
"module_path": "olive.passes.onnx.mixed_precision.OrtMixedPrecision",
"supported_providers": [ "CUDAExecutionProvider", "DmlExecutionProvider" ],
"supported_accelerators": [ "gpu", "npu" ],
"supported_precisions": [ "fp16" ]
},
"OrtSessionParamsTuning": {
"module_path": "olive.passes.onnx.session_params_tuning.OrtSessionParamsTuning",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"module_dependencies": [ "psutil" ]
},
"OrtTransformersOptimization": {
"module_path": "olive.passes.onnx.transformer_optimization.OrtTransformersOptimization",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"QNNPreprocess": {
"module_path": "olive.passes.onnx.qnn.qnn_preprocess.QNNPreprocess",
"supported_providers": [ "QNNExecutionProvider" ],
"supported_accelerators": [ "npu" ],
"supported_precisions": [ "*" ]
},
"VitisAIQuantization": {
"module_path": "olive.passes.onnx.vitis_ai_quantization.VitisAIQuantization",
"supported_providers": [ "VitisAIExecutionProvider" ],
"supported_accelerators": [ "npu" ],
"supported_precisions": [ "int8" ]
},
"VitisQDQQuantizer": {
"module_path": "olive.passes.onnx.vitis_ai.quantizer.VitisQDQQuantizer",
"supported_providers": [ "VitisAIExecutionProvider" ],
"supported_accelerators": [ "npu" ],
"supported_precisions": [ "int8" ]
},
"VitisQOpQuantizer": {
"module_path": "olive.passes.onnx.vitis_ai.quantizer.VitisQOpQuantizer",
"supported_providers": [ "VitisAIExecutionProvider" ],
"supported_accelerators": [ "npu" ],
"supported_precisions": [ "int8" ]
},
"PowerOfTwoMethod": {
"module_path": "olive.passes.onnx.vitis_ai.quant_utils.PowerOfTwoMethod",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"SplitModel": {
"module_path": "olive.passes.onnx.split.SplitModel",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"OpenVINOConversion": {
"module_path": "olive.passes.openvino.conversion.OpenVINOConversion",
"supported_providers": [ "OpenVINOExecutionProvider" ],
"supported_accelerators": [ "cpu", "gpu", "npu" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "openvino" ]
},
"OpenVINOQuantization": {
"module_path": "olive.passes.openvino.quantization.OpenVINOQuantization",
"supported_providers": [ "OpenVINOExecutionProvider" ],
"supported_accelerators": [ "cpu", "gpu", "npu" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "openvino" ]
},
"AutoAWQQuantizer": {
"module_path": "olive.passes.pytorch.autoawq.AutoAWQQuantizer",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int4", "int8", "int16", "uint4", "uint8", "uint16" ],
"module_dependencies": [ "autoawq" ]
},
"GptqQuantizer": {
"module_path": "olive.passes.pytorch.gptq.GptqQuantizer",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int4", "int8", "int16", "uint4", "uint8", "uint16" ],
"module_dependencies": [ "auto-gptq", "optimum" ]
},
"CaptureSplitInfo": {
"module_path": "olive.passes.pytorch.capture_split_info.CaptureSplitInfo",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"MergeAdapterWeights": {
"module_path": "olive.passes.pytorch.merge_adapter_weights.MergeAdapterWeights",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"LoftQ": {
"module_path": "olive.passes.pytorch.lora.LoftQ",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"LoRA": {
"module_path": "olive.passes.pytorch.lora.LoRA",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "lora" ]
},
"PyTorchTensorParallel": {
"module_path": "olive.passes.pytorch.tensor_parallel.PyTorchTensorParallel",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"QLoRA": {
"module_path": "olive.passes.pytorch.lora.QLoRA",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "bnb", "lora" ]
},
"QuantizationAwareTraining": {
"module_path": "olive.passes.pytorch.quantization_aware_training.QuantizationAwareTraining",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"module_dependencies": [ "pytorch-lightning" ]
},
"SparseGPT": {
"module_path": "olive.passes.pytorch.sparsegpt.SparseGPT",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"SliceGPT": {
"module_path": "olive.passes.pytorch.slicegpt.SliceGPT",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"QuaRot": {
"module_path": "olive.passes.pytorch.quarot.QuaRot",
"supported_providers": [ "CPUExecutionProvider" ],
"supported_accelerators": [ "cpu" ],
"supported_precisions": [ "int4", "int8", "int16", "uint4", "uint8", "uint16" ],
"extra_dependencies": [ "flash-attn" ]
},
"TorchTRTConversion": {
"module_path": "olive.passes.pytorch.torch_trt_conversion.TorchTRTConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ],
"extra_dependencies": [ "torch-tensorrt" ]
},
"QNNConversion": {
"module_path": "olive.passes.qnn.conversion.QNNConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"QNNModelLibGenerator": {
"module_path": "olive.passes.qnn.model_lib_generator.QNNModelLibGenerator",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"QNNContextBinaryGenerator": {
"module_path": "olive.passes.qnn.context_binary_generator.QNNContextBinaryGenerator",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"SNPEConversion": {
"module_path": "olive.passes.snpe.conversion.SNPEConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"SNPEQuantization": {
"module_path": "olive.passes.snpe.quantization.SNPEQuantization",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"SNPEtoONNXConversion": {
"module_path": "olive.passes.snpe.snpe_to_onnx.SNPEtoONNXConversion",
"supported_providers": [ "*" ],
"supported_accelerators": [ "*" ],
"supported_precisions": [ "*" ]
},
"NVModelOptQuantization": {
"module_path": "olive.passes.onnx.nvmo_quantization.NVModelOptQuantization",
"supported_providers": [ "CUDAExecutionProvider" ],
"supported_accelerators": [ "gpu" ],
"supported_precisions": [ "int4", "int8", "fp8" ],
"extra_dependencies": [ "nvmo" ]
}
},
"extra_dependencies": {
"auto-opt": [ "optimum" ],
"azureml": [ "azure-ai-ml>=1.11.1", "azure-keyvault-secrets", "azure-identity", "azureml-fsspec" ],
"bnb": [ "bitsandbytes" ],
"capture-onnx-graph": [ "onnxruntime-genai", "optimum" ],
"cpu": [ "onnxruntime" ],
"directml": [ "onnxruntime-directml" ],
"docker": [ "docker" ],
"shared-cache": [ "azure-identity", "azure-storage-blob" ],
"finetune": [ "onnxruntime-genai", "optimum", "accelerate>=0.30.0", "peft", "scipy", "bitsandbytes" ],
"flash-attn": [ "flash_attn" ],
"gpu": [ "onnxruntime-gpu" ],
"inc": [ "neural-compressor" ],
"lora": [ "accelerate>=0.30.0", "peft", "scipy" ],
"nvmo": [ "nvidia-modelopt", "onnx-graphsurgeon", "datasets>=2.14.4" ],
"openvino": [ "openvino==2023.2.0", "nncf==2.7.0", "numpy<2.0" ],
"optimum": [ "optimum" ],
"ort-genai": [ "onnxruntime-genai" ],
"ort": [ "onnxruntime", "onnxruntime-directml", "onnxruntime-gpu", "onnxruntime-openvino", "numpy<2.0" ],
"tf": [ "tensorflow==1.15.0" ],
"torch-tensorrt": [ "torch-tensorrt" ],
"tune-session-params": [ "psutil" ]
}
}