Skip to content

Update dynamic quant tutorial for saving quantized model and accurate numerics #1167

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 25, 2020
Merged

Update dynamic quant tutorial for saving quantized model and accurate numerics #1167

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 14 additions & 6 deletions intermediate_source/dynamic_quantization_bert_tutorial.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -492,7 +492,7 @@ follows:

| Prec | F1 score | Model Size | 1 thread | 4 threads |
| FP32 | 0.9019 | 438 MB | 160 sec | 85 sec |
| INT8 | 0.8953 | 181 MB | 90 sec | 46 sec |
| INT8 | 0.902 | 181 MB | 90 sec | 46 sec |

We have 0.6% F1 score accuracy after applying the post-training dynamic
quantization on the fine-tuned BERT model on the MRPC task. As a
Expand Down Expand Up @@ -520,15 +520,23 @@ processing the evaluation of MRPC dataset.
3.3 Serialize the quantized model
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

We can serialize and save the quantized model for the future use.
We can serialize and save the quantized model for the future use using
`torch.jit.save` after tracing the model.

.. code:: python

quantized_output_dir = configs.output_dir + "quantized/"
if not os.path.exists(quantized_output_dir):
os.makedirs(quantized_output_dir)
quantized_model.save_pretrained(quantized_output_dir)
input_ids = ids_tensor([8, 128], 2)
token_type_ids = ids_tensor([8, 128], 2)
attention_mask = ids_tensor([8, 128], vocab_size=2)
dummy_input = (input_ids, attention_mask, token_type_ids)
traced_model = torch.jit.trace(quantized_model, dummy_input)
torch.jit.save(traced_model, "bert_traced_eager_quant.pt")

To load the quantized model, we can use `torch.jit.load`

.. code:: python

loaded_quantized_model = torch.jit.load("bert_traced_eager_quant.pt")

Conclusion
----------
Expand Down