Skip to content

GitLab

"examples/vscode:/vscode.git/clone" did not exist on "ba2ca4be6d20f9e661d396a8b3b76182b960cfc5"
Switch branch/tag
  1. 08 Jul, 2022 2 commits
    • Yanghan Wang's avatar
      prepare_for_quant_convert -> custom_covert_fx · 97904ba4
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/325

`prepare_for_quant_convert` is a confusing name because it only does `convert`, there's no "prepare" in it. It's actually for fx only, because eager mode always calls `torch.quantization.convert`, there's no way to customize it. So just call this `custom_convert_fx`. The new name is also unique in fbcode, so easy to do codemod later on.

This diff simply does the renaming by biggrep + replace.

Reviewed By: jerryzh168

Differential Revision: D37676717

fbshipit-source-id: e7d05eaafddc383dd432986267c945c8ebf94df4
      97904ba4
    • Yanghan Wang's avatar
      fix FX mode rcnn avgpool output config · c7226783
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/327

Fix the bug for not leaving avgpool quantized when its input is quantized. This is not captured by CI because currently we set `small_pooler_resolution` to True to make test running faster, but this skips the `avgpool` everytime.

Reviewed By: jerryzh168

Differential Revision: D37706151

fbshipit-source-id: fde9a190324a23b97ba803e000f4caa54f6f3738
      c7226783
  3. 06 Jul, 2022 2 commits
  5. 02 Jul, 2022 1 commit
  7. 30 Jun, 2022 2 commits
  9. 29 Jun, 2022 2 commits
  11. 28 Jun, 2022 2 commits
  13. 27 Jun, 2022 1 commit
  15. 24 Jun, 2022 3 commits
    • Yanghan Wang's avatar
      better organize builtin runner's default config · 07eb7a1e
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/313

It's not natural to put runner's default config functions under `d2go/utils/`, move them to `d2go/runner/config_defaults.py` and clean things up. This also reduce the inter-sub-package dependencies.

Reviewed By: mattcyu1

Differential Revision: D37407078

fbshipit-source-id: 432644bee4f12306a14bac3dba76ced08b3683aa
      07eb7a1e
    • Mik Vyatskov's avatar
      Only save results to file from rank 0 · f0297b81
      Mik Vyatskov authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/309

Right now multiple machines can try to write to the same output file,
since they get the same argument. Additionally, on the same machine, several
outputs can be saved which requires unncessary unpacking. This change makes
train_net only write output of the rank 0 trainer.

Reviewed By: wat3rBro

Differential Revision: D37310084

fbshipit-source-id: 9d5352a274e8fb1d2043393b12896d402333c17b
      f0297b81
    • Yanghan Wang's avatar
      use runner class instead of instance outside of main · 8051775c
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/312

As discussed, we decided to not use runner instance outside of `main`, previous diffs already solved the prerequisites, this diff mainly does the renaming.
- Use runner name (str) in the fblearner, ML pipeline.
- Use runner name (str) in FBL operator, MAST and binary operator.
- Use runner class as the interface of main, it can be either the name of class (str) or actual class. The main usage should be using `str`, so that the importing of class happens inside `main`. But it's also a common use case to import runner class and call `main` for things like ad-hoc scripts or tests, supporting actual class makes it easier modify code for those cases (eg. some local test class doesn't have a name, so it's not feasible to use runner name).

Reviewed By: newstzpz

Differential Revision: D37060338

fbshipit-source-id: 879852d41902b87d6db6cb9d7b3e8dc55dc4b976
      8051775c
  17. 22 Jun, 2022 2 commits
  19. 21 Jun, 2022 2 commits
  21. 20 Jun, 2022 1 commit
  23. 18 Jun, 2022 2 commits
  25. 17 Jun, 2022 3 commits
  27. 16 Jun, 2022 8 commits
    • Yanghan Wang's avatar
      support for using config file with _DEFAULTS_ via cli · 521b3cad
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/271

- set `get_default_cfg = None` to opt-in this new feature.
- support config with `_DEFAULTS_` when loading config file. Note that we don't check explicitly `"_DEFAULTS_" in config file` but check `runner.get_default_cfg == None`, this is because runner having `get_default_cfg` and config having `_DEFAULTS_` should be mutually exclusive, and `load_full_config_from_file` can raise proper error if `_DEFAULTS_` is missing.
- we also need to save `_DEFAULTS_` in the diff config.

Reviewed By: tglik

Differential Revision: D36868581

fbshipit-source-id: e0e19309c3df5a85383ce1454b321a68d0868dc4
      521b3cad
    • Yanghan Wang's avatar
      restructure lightning related code · 318a3d79
      Yanghan Wang authored 
      Summary: Pull Request resolved: https://github.com/facebookresearch/d2go/pull/237

Reviewed By: tglik

Differential Revision: D35954531

fbshipit-source-id: b69c8065928fe385d29f20f2c2460d60d63fca00
      318a3d79
    • Matthew Yu's avatar
      add modeling hook algo and helper · f3fc01aa
      Matthew Yu authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/299

This implements the first iteration of generalized distillation in D2 (https://github.com/facebookresearch/d2go/commit/87374efb134e539090e0b5c476809dc35bf6aedb)Go. The functionality is separated into the following:

=> Adding distillation functionality without user changing their meta architecture:

class DistillationModelingHook
* This is an implementation detail that we hide from the user.
* We use dynamic mixin to specify additional functionality to the user's model. In this way, the original (student) model retains all attributes but the mixin class will override the forward (and provide more functionality like teacher updates).
* Teacher build currently only supports loading a torchscript model, pytorch compatiblity in later diffs

=> Implementing distillation methods

class DistillationAlgorithm
* The user can use some default algorithm (e.g., LabelDistillation) or create their own. This is where we specify the overrided forward func of the model and any other distillation requirements (updating the weights of the teacher model).
* The basic LabelDistillation allows a user to use a teacher model during training to relabel the gt

=> User customization

class DistillationHelper
* This is what we expect the user to customize. As an example the user probably needs to write their own pseudo_labeler to take batched_inputs and relabel this with the teacher

Both DistillationHelper and DistillationAlgorithm use a registry so that users can add their customization in their own code and use these customizations by specifying in the config

Reviewed By: newstzpz

Differential Revision: D36708227

fbshipit-source-id: bc427c5d42d0c7ff4d839bf10782efac24dea107
      f3fc01aa
    • Miquel Jubert Hermoso's avatar
      Hardcode top-level directory, remove import for resource access · ce776f7d
      Miquel Jubert Hermoso authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/229

*This diff is part of a stack which has the goal of "buckifying" D2 (https://github.com/facebookresearch/d2go/commit/87374efb134e539090e0b5c476809dc35bf6aedb)Go core and enabling autodeps and other tooling. The last diff in the stack introduces the TARGETS. The diffs earlier in the stack are resolving circular dependencies and other issues which prevent the buckification from occurring.*

This diff changes the import paths being rerouted in the d2go package. Instead of local-importing the package, which creates a buck circular dependency, we can hardcode the top-level path. We know it is fixed because we are defining it with base_module.

For the other packages it can remain as-is.

Also add some dependencies annotations in preparation for the buckification, with manual.

Reviewed By: wat3rBro

Differential Revision: D35928513

fbshipit-source-id: 756c95740bab519e8093d254bcee64692e1a1b98
      ce776f7d
    • Mik Vyatskov's avatar
      Implement a central helper for converting arguments to CLI args · 3b64e76a
      Mik Vyatskov authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/301

This is a follow-up of earlier work to extract part responsible for the
centrally defined parameters from the helper in train_net closer to where the
parameters are defined.

Reviewed By: tglik

Differential Revision: D37176212

fbshipit-source-id: 226415f36f4872ac3d9ba41541b4389a18cc11e6
      3b64e76a
    • Mircea Cimpoi's avatar
      Refactor quantization logic to separate function · 5654d831
      Mircea Cimpoi authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/296

Extract the model quantization logic to separate function.
+minor cleanup (remove unused imports, f-strings)

Reviewed By: wat3rBro

Differential Revision: D36999434

fbshipit-source-id: 7aad64921b8cdf8779527c19b077ee788403c6b8
      5654d831
    • Yanghan Wang's avatar
      make launch return results from all ranks · 55dc3da1
      Yanghan Wang authored 
      Summary:
X-link: https://github.com/facebookresearch/mobile-vision/pull/86

Pull Request resolved: https://github.com/facebookresearch/d2go/pull/300

Similar to `elastic_launch`, make `launch` return results for all ranks, this is beneficial for cases where each rank having different return value`.

There'll be some test failing for V1, it'll help finding related callsites.

Reviewed By: tglik

Differential Revision: D37016935

fbshipit-source-id: eb989342c0ddd642c8369b5dddfa8aac8cb38917
      55dc3da1
    • Shiyu Dong's avatar
      Add CallFuncWithNameAndJsonFile · d6a966fd
      Shiyu Dong authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/288

Add a new call function to handle adhoc datasets that needs to be called with dataset_name. This is mainly to handle the incompatibility of `extend_coco_load` function with AdhocDataset

For example,
* if we choose to register `extend_coco_load` function as CallFuncWithJsonFile, AdhocDataset will register the same function, while failed to update the `dataset_name` argument passed to `extended_coco_load`.

* If we choose to register `extend_coco_load` function normally, we couldn't use the `LOAD_KWARGS` that specifies additional fields like `image_direct_copy_keys` when register AdhocDatasets

Therefore adding the new function allows to update register function with specifying new json as well as new dataset_name in the same time.

Reviewed By: sstsai-adl

Differential Revision: D37100742

fbshipit-source-id: 93f4a7ac23f95812aca80c40872b15f92d449da0
      d6a966fd
  29. 15 Jun, 2022 4 commits
  31. 14 Jun, 2022 3 commits
    • Yanghan Wang's avatar
      make get_default_cfg a classmethod · 65dad512
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/293

In order to pass runner during the workflow using "runner name" instead of runner instance, we need to make sure the `get_default_cfg` is not instance method. It can be either staticmethod or classmethod, but I choose classmethod for better inheritance.

code mode using following script:
```
#!/usr/bin/env python3

import json
import os
import subprocess

result = subprocess.check_output("fbgs --json 'def get_default_cfg('", shell=True)
fbgs = json.loads(result)
fbsource_root = os.path.expanduser("~")

def _indent(s):
    return len(s) - len(s.lstrip())

def resolve_instance_method(content):
    lines = content.split("\n")
    for idx, line in enumerate(lines):
        if "def get_default_cfg(self" in line:
            indent = _indent(line)
            # find the class
            for j in range(idx, 0, -1):
                if lines[j].startswith(" " * (indent - 4) + "class "):
                    class_line = lines[j]
                    break
            else:
                raise RuntimeError("Can't find class")
            print("class_line: ", class_line)
            if "Runner" in class_line:
                # check self if not used
                for j in range(idx + 1, len(lines)):
                    if _indent(lines[j]) < indent:
                        break
                    assert "self" not in lines[j], (j, lines[j])
                # update the content
                assert "def get_default_cfg(self)" in line
                lines[idx] = lines[idx].replace(
                    "def get_default_cfg(self)", "def get_default_cfg(cls)"
                )
                lines.insert(idx, " " * indent + "classmethod")
                return "\n".join(lines)
    return content

def resolve_static_method(content):
    lines = content.split("\n")
    for idx, line in enumerate(lines):
        if "def get_default_cfg()" in line:
            indent = _indent(line)
            # find the class
            for j in range(idx, 0, -1):
                if "class " in lines[j]:
                    class_line = lines[j]
                    break
            else:
                print("[WARNING] Can't find class!!!")
                continue
            if "Runner" in class_line:
                # check staticmethod is used
                for j in range(idx, 0, -1):
                    if lines[j] == " " * indent + "staticmethod":
                        staticmethod_line_idx = j
                        break
                else:
                    raise RuntimeError("Can't find staticmethod")
                # update the content
                lines[idx] = lines[idx].replace(
                    "def get_default_cfg()", "def get_default_cfg(cls)"
                )
                lines[staticmethod_line_idx] = " " * indent + "classmethod"
                return "\n".join(lines)
    return content

for result in fbgs["results"]:
    filename = os.path.join(fbsource_root, result["file_name"])
    print(f"processing: {filename}")
    with open(filename) as f:
        content = f.read()
    orig_content = content
    while True:
        old_content = content
        content = resolve_instance_method(content)
        content = resolve_static_method(content)
        if content == old_content:
            break
    if content != orig_content:
        print("Updating ...")
        with open(filename, "w") as f:
            f.write(content)
```

Reviewed By: tglik

Differential Revision: D37059264

fbshipit-source-id: b09d5518f4232de95d8313621468905cf10a731c
      65dad512
    • Yanghan Wang's avatar
      support diff config for lightning_train_net · 8cf2b879
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/285

`setup_after_launch` can now take `DefaultTask` as well (the `runner_or_task` can still be `None`, for runner-less train_net).

Reviewed By: tglik

Differential Revision: D37011560

fbshipit-source-id: ce8a88242df0a16de8da97d94e8eb7def524c69c
      8cf2b879
    • Sam Tsai's avatar
      separating out save function and checking if it has been run previously for a dataset · 84dac84f
      Sam Tsai authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/282

Writing the of the converted coco file needs to be synced since the file is used by COCOEval. The whole function could be called multiple times by main thread so need to add additional checks to see if dataset has been previous called.

Reviewed By: wat3rBro

Differential Revision: D36957778

fbshipit-source-id: 5f9c5c038c13cbe24132ef1b44f6e94d5a26b66a
      84dac84f