Onnx transformers: Quantize option #6
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
-changed framework type to "pt"
added usage example with Ner pipe.
reading configuration from model config
patil-suraj
requested changes
Oct 13, 2020
| modelcard = config | ||
| # TODO: Disable modelcard (below 4 lines) if working with local models. | ||
| # searches modelcard.json | ||
| if not local_model: |
There was a problem hiding this comment.
we can keep it as it is, does it break for local model ? if not let's keep it as it is
| @@ -694,6 +704,20 @@ def _forward(self, inputs, return_tensors=False): | |||
| else: | |||
| return predictions.numpy() | |||
|
|
|||
| def _create_quantized_graph(self, onnx_opt_model_path): | |||
| #TODO: add option gpt2 if need | |||
| opt_options = BertOptimizationOptions('bert') | |||
There was a problem hiding this comment.
check model type explicitly, raise an assert or exception if the model is not bert
| @@ -555,7 +559,13 @@ def __init__( | |||
| logger.info(f"loading onnx graph from {self.graph_path.as_posix()}") | |||
| self.onnx_model = create_model_for_provider(str(graph_path), "CPUExecutionProvider") | |||
| self.input_names = json.load(open(input_names_path)) | |||
| self.framework = "np" | |||
| self.framework = "pt" | |||
There was a problem hiding this comment.
This will cause other things to break. Are all tests passing ?
you can run tests using make test.
| Set `onnx` to `False` for standard torch inference. | ||
| Set `quantized` to `True` for quantize with Onnx. ( set `onnx` to True) |
There was a problem hiding this comment.
Suggested change
| Set `quantized` to `True` for quantize with Onnx. ( set `onnx` to True) | |
| Set `quantized` to `True` for quantize with Onnx. ( set `onnx` to True) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I've made changes as we talked in pr.
I added option
local_modelto pipeline. It ignoresmodelcardto load local models that without having modelcard.I kept framework as torch. In some cases like loading local models i have got error that
InvalidArgument: [ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Unexpected input data type.. We can leave it like this to stay safe.