diff --git a/examples/recipes/README.md b/examples/recipes/README.md index 86aa01869..77606ae03 100644 --- a/examples/recipes/README.md +++ b/examples/recipes/README.md @@ -14,7 +14,7 @@ Each *(model, task)* includes: ## Models -Total: **75** (model, task) tuples that pass fp16 eval on all 10 (EP, device) buckets. +Total: **76** (model, task) tuples that pass fp16 eval on all 10 (EP, device) buckets. | Model | Task | |---|---| @@ -43,6 +43,7 @@ Total: **75** (model, task) tuples that pass fp16 eval on all 10 (EP, device) bu | deepset/bert-large-uncased-whole-word-masking-squad2 | question-answering | | deepset/roberta-base-squad2 | question-answering | | deepset/tinyroberta-squad2 | question-answering | +| dell-research-harvard/lt-wikidata-comp-zh | feature-extraction | | dima806/fairface_age_image_detection | image-classification | | distilbert/distilbert-base-cased-distilled-squad | question-answering | | distilbert/distilbert-base-uncased | fill-mask | diff --git a/examples/recipes/dell-research-harvard_lt-wikidata-comp-zh/feature-extraction_fp16_config.json b/examples/recipes/dell-research-harvard_lt-wikidata-comp-zh/feature-extraction_fp16_config.json new file mode 100644 index 000000000..787760f0d --- /dev/null +++ b/examples/recipes/dell-research-harvard_lt-wikidata-comp-zh/feature-extraction_fp16_config.json @@ -0,0 +1,101 @@ +{ + "export": { + "opset_version": 17, + "batch_size": 1, + "export_params": true, + "do_constant_folding": true, + "verbose": false, + "dynamo": false, + "enable_hierarchy_tags": true, + "clean_onnx": false, + "hierarchy_tag_format": "full", + "input_tensors": [ + { + "name": "input_ids", + "dtype": "int32", + "shape": [ + 1, + 512 + ], + "value_range": [ + 0, + 21128 + ] + }, + { + "name": "attention_mask", + "dtype": "int32", + "shape": [ + 1, + 512 + ], + "value_range": [ + 0, + 2 + ] + }, + { + "name": "token_type_ids", + "dtype": "int32", + "shape": [ + 1, + 512 + ], + "value_range": [ + 0, + 2 + ] + } + ], + "output_tensors": [ + { + "name": "last_hidden_state" + } + ] + }, + "optim": { + "clamp_constant_values": true + }, + "quant": { + "mode": "fp16", + "samples": 10, + "calibration_method": "minmax", + "weight_type": "uint8", + "activation_type": "uint8", + "per_channel": false, + "symmetric": false, + "weight_symmetric": null, + "activation_symmetric": null, + "save_calibration": false, + "distribution": "uniform", + "seed": null, + "calibration_load_path": null, + "calibration_save_path": null, + "op_types_to_quantize": null, + "nodes_to_exclude": null, + "task": "feature-extraction", + "model_id": "dell-research-harvard/lt-wikidata-comp-zh", + "model_type": "bert", + "fp16_keep_io_types": true, + "fp16_op_block_list": null + }, + "compile": null, + "loader": { + "task": "feature-extraction", + "model_class": "AutoModel", + "model_type": "bert" + }, + "eval": { + "task": "feature-extraction", + "dataset": { + "path": "mteb/stsbenchmark-sts", + "split": "test", + "samples": 100, + "columns_mapping": { + "input_column_1": "sentence1", + "input_column_2": "sentence2", + "score_column": "score" + } + } + } +} diff --git a/examples/recipes/dell-research-harvard_lt-wikidata-comp-zh/feature-extraction_w8a16_config.json b/examples/recipes/dell-research-harvard_lt-wikidata-comp-zh/feature-extraction_w8a16_config.json new file mode 100644 index 000000000..c5519d6be --- /dev/null +++ b/examples/recipes/dell-research-harvard_lt-wikidata-comp-zh/feature-extraction_w8a16_config.json @@ -0,0 +1,99 @@ +{ + "export": { + "opset_version": 17, + "batch_size": 1, + "export_params": true, + "do_constant_folding": true, + "verbose": false, + "dynamo": false, + "enable_hierarchy_tags": true, + "clean_onnx": false, + "hierarchy_tag_format": "full", + "input_tensors": [ + { + "name": "input_ids", + "dtype": "int32", + "shape": [ + 1, + 512 + ], + "value_range": [ + 0, + 21128 + ] + }, + { + "name": "attention_mask", + "dtype": "int32", + "shape": [ + 1, + 512 + ], + "value_range": [ + 0, + 2 + ] + }, + { + "name": "token_type_ids", + "dtype": "int32", + "shape": [ + 1, + 512 + ], + "value_range": [ + 0, + 2 + ] + } + ], + "output_tensors": [ + { + "name": "last_hidden_state" + } + ] + }, + "optim": { + "clamp_constant_values": true + }, + "quant": { + "mode": "static", + "samples": 10, + "calibration_method": "minmax", + "weight_type": "uint8", + "activation_type": "uint16", + "per_channel": false, + "symmetric": false, + "weight_symmetric": null, + "activation_symmetric": null, + "save_calibration": false, + "distribution": "uniform", + "seed": null, + "calibration_load_path": null, + "calibration_save_path": null, + "op_types_to_quantize": null, + "nodes_to_exclude": null, + "task": "feature-extraction", + "model_id": "dell-research-harvard/lt-wikidata-comp-zh", + "model_type": "bert" + }, + "compile": null, + "loader": { + "task": "feature-extraction", + "model_class": "AutoModel", + "model_type": "bert" + }, + "eval": { + "task": "feature-extraction", + "dataset": { + "path": "mteb/stsbenchmark-sts", + "split": "test", + "samples": 1000, + "columns_mapping": { + "input_column_1": "sentence1", + "input_column_2": "sentence2", + "score_column": "score" + } + } + } +}