SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Model debugger upgrades (#37391)

Browse Source 
* debugging improvements

* add debugging details

* add more debugging details

* debug more

* clean up layers + output

* add summary json file

* cleanup

* copies 👀

* remove hooks + add documentation

* draft a small test, why not

* respect the format (respect it)

* fixup imports

* nit

* add tests and configurable pruning of layers
This commit is contained in:
Pablo Montalvo
2025-04-18 16:45:54 +02:00
committed by GitHub
parent e5ac23081e
commit 4afd3f4820
5 changed files with 425 additions and 112 deletions
Download Patch File Download Diff File Expand all files Collapse all files

148
docs/source/en/internal/model_debugging_utils.md
View File

@@ -28,7 +28,7 @@ Most of those are only useful if you are adding new models in the library.
This context manager is a power user tool intended for model adders. This context manager is a power user tool intended for model adders.
It tracks all forward calls within a model forward and logs a slice of each input and output on a nested Json. It tracks all forward calls within a model forward and logs a slice of each input and output on a nested Json.
To note, this context manager enforces `torch.inference_mode()`. To note, this context manager enforces `torch.no_grad()`.
### Rationale ### Rationale
@@ -43,6 +43,7 @@ import torch
from PIL import Image from PIL import Image
import requests import requests
from transformers import LlavaProcessor, LlavaForConditionalGeneration from transformers import LlavaProcessor, LlavaForConditionalGeneration
from transformers.model_debugging_utils import model_addition_debugger_context
torch.random.manual_seed(673) torch.random.manual_seed(673)
# load pretrained model and processor # load pretrained model and processor
@@ -60,12 +61,153 @@ prompt = "<image>Describe this image."
inputs = processor(text=prompt, images=random_image, return_tensors="pt") inputs = processor(text=prompt, images=random_image, return_tensors="pt")
# call forward method (not .generate!) # call forward method (not .generate!)
with model_addition_debugger_context(model, "optional_path_to_your_output_file.json"): with model_addition_debugger_context(
model,
debug_path="optional_path_to_your_directory",
do_prune_layers=False # This will output ALL the layers of a model.
):
output = model.forward(**inputs) output = model.forward(**inputs)
``` ```
[[autodoc]] model_addition_debugger ### Reading results
The debugger generates two files from the forward call, both with the same base name,
but ending either with `_SUMMARY.json` or with `_FULL_TENSORS.json`.
The first one will contain a summary of each module's _input_ and _output_ tensor values and shapes.
```json
{
"module_path": "MolmoForConditionalGeneration",
"inputs": {
"args": [],
"kwargs": {
"input_ids": {
"shape": "torch.Size([1, 589])",
"dtype": "torch.int64"
},
"attention_mask": {
"shape": "torch.Size([1, 589])",
"dtype": "torch.int64"
},
"pixel_values": {
"shape": "torch.Size([1, 5, 576, 588])",
"dtype": "torch.float32",
"mean": "tensor(-8.9514e-01, device='cuda:0')",
"std": "tensor(9.2586e-01, device='cuda:0')",
"min": "tensor(-1.7923e+00, device='cuda:0')",
"max": "tensor(1.8899e+00, device='cuda:0')"
}
},
"children": [
{
"module_path": "MolmoForConditionalGeneration.language_model.model.embed_tokens",
"inputs": {
"args": [
{
"shape": "torch.Size([1, 589])",
"dtype": "torch.int64"
}
]
},
"outputs": {
"shape": "torch.Size([1, 589, 3584])",
"dtype": "torch.float32",
"mean": "tensor(6.5460e-06, device='cuda:0')",
"std": "tensor(2.3807e-02, device='cuda:0')",
"min": "tensor(-3.3398e-01, device='cuda:0')",
"max": "tensor(3.9453e-01, device='cuda:0')"
}
},
{
"module_path": "MolmoForConditionalGeneration.vision_tower",
"inputs": {
"args": [
{
"shape": "torch.Size([5, 1, 576, 588])",
"dtype": "torch.float32",
"mean": "tensor(-8.9514e-01, device='cuda:0')",
"std": "tensor(9.2586e-01, device='cuda:0')",
"min": "tensor(-1.7923e+00, device='cuda:0')",
"max": "tensor(1.8899e+00, device='cuda:0')"
}
],
"kwargs": {
"output_hidden_states": "True"
}
},
"children": [
{ ... and so on
```
The `_FULL_TENSORS.json` file will display a full view of all tensors, which is useful
for comparing two files.
```json
"pixel_values": {
"shape": "torch.Size([1, 5, 576, 588])",
"dtype": "torch.float32",
"value": [
"tensor([[[[-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" ...,",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00]],",
"",
" [[-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" ...,",
" [-1.4857e+00, -1.4820e+00, -1.2100e+00, ..., -6.0979e-01, -5.9650e-01, -3.8527e-01],",
" [-1.6755e+00, -1.7221e+00, -1.4518e+00, ..., -7.5577e-01, -7.4658e-01, -5.5592e-01],",
" [-7.9957e-01, -8.2162e-01, -5.7014e-01, ..., -1.3689e+00, -1.3169e+00, -1.0678e+00]],",
"",
" [[-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" ...,",
" [-3.0322e-01, -5.0645e-01, -5.8436e-01, ..., -6.2439e-01, -7.9160e-01, -8.1188e-01],",
" [-4.4921e-01, -6.5653e-01, -7.2656e-01, ..., -3.4702e-01, -5.2146e-01, -5.1326e-01],",
" [-3.4702e-01, -5.3647e-01, -5.4170e-01, ..., -1.0915e+00, -1.1968e+00, -1.0252e+00]],",
"",
" [[-1.1207e+00, -1.2718e+00, -1.0678e+00, ..., 1.2013e-01, -1.3126e-01, -1.7197e-01],",
" [-6.9738e-01, -9.1166e-01, -8.5454e-01, ..., -5.5050e-02, -2.8134e-01, -4.2793e-01],",
" [-3.4702e-01, -5.5148e-01, -5.8436e-01, ..., 1.9312e-01, -8.6235e-02, -2.1463e-01],",
" ...,",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00]],",
"",
" [[-1.0039e+00, -9.5669e-01, -6.5546e-01, ..., -1.4711e+00, -1.4219e+00, -1.1389e+00],",
" [-1.0039e+00, -9.5669e-01, -6.5546e-01, ..., -1.7193e+00, -1.6771e+00, -1.4091e+00],",
" [-1.6317e+00, -1.6020e+00, -1.2669e+00, ..., -1.2667e+00, -1.2268e+00, -8.9720e-01],",
" ...,",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00],",
" [-1.7923e+00, -1.7521e+00, -1.4802e+00, ..., -1.7923e+00, -1.7521e+00, -1.4802e+00]]]], device='cuda:0')"
],
"mean": "tensor(-8.9514e-01, device='cuda:0')",
"std": "tensor(9.2586e-01, device='cuda:0')",
"min": "tensor(-1.7923e+00, device='cuda:0')",
"max": "tensor(1.8899e+00, device='cuda:0')"
},
```
### Comparing between implementations
Once the forward passes of two models have been traced by the debugger, one can compare the `json` output files. See below: we can see slight differences between these two implementations' key projection layer. Inputs are mostly identical, but not quite. Looking through the file differences makes it easier to pinpoint which layer is wrong.
![download-icon](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/files_difference_debugging.png)
### Limitations and scope
This feature will only work for torch-based models, and would require more work and case-by-case approach for say `jax`-based models that are usually compiled. Models relying heavily on external kernel calls may work, but trace will probably miss some things. Regardless, any python implementation that aims at mimicking another implementation can be traced once instead of reran N times with breakpoints.
If you pass `do_prune_layers=False` to your model debugger, ALL the layers will be outputted to `json`. Else, only the first and last layer will be shown. This is useful when some layers (typically cross-attention) appear only after N layers.
[[autodoc]] model_addition_debugger_context [[autodoc]] model_addition_debugger_context

2
src/transformers/__init__.py
View File

@@ -344,7 +344,6 @@ except OptionalDependencyNotAvailable:
_import_structure["utils.dummy_pt_objects"] = [name for name in dir(dummy_pt_objects) if not name.startswith("_")] _import_structure["utils.dummy_pt_objects"] = [name for name in dir(dummy_pt_objects) if not name.startswith("_")]
else: else:
_import_structure["model_debugging_utils"] = [ _import_structure["model_debugging_utils"] = [
"model_addition_debugger",
"model_addition_debugger_context", "model_addition_debugger_context",
] ]
_import_structure["activations"] = [] _import_structure["activations"] = []
@@ -910,7 +909,6 @@ if TYPE_CHECKING:
convert_and_export_with_cache, convert_and_export_with_cache,
) )
from .model_debugging_utils import ( from .model_debugging_utils import (
model_addition_debugger,
model_addition_debugger_context, model_addition_debugger_context,
) )
from .modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update from .modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update

259
src/transformers/model_debugging_utils.py
View File

@@ -17,7 +17,8 @@ import functools
import json import json
import os import os
import re import re
from contextlib import contextmanager from contextlib import contextmanager, redirect_stdout
from io import StringIO
from typing import Optional from typing import Optional
from transformers.utils.import_utils import requires from transformers.utils.import_utils import requires
@@ -28,9 +29,7 @@ from .utils import is_torch_available
if is_torch_available(): if is_torch_available():
import torch import torch
import torch.distributed.tensor import torch.distributed.tensor
from torch import nn
from .modeling_utils import PreTrainedModel
from .utils import logging from .utils import logging
@@ -87,21 +86,64 @@ def _serialize_io(value):
if hasattr(value, "_local_tensor"): if hasattr(value, "_local_tensor"):
# DTensor-like handling, just use local tensor attribute # DTensor-like handling, just use local tensor attribute
return { torch.set_printoptions(sci_mode=True)
val_repr = _repr_to_list(value)
out = {
"shape": repr(value._local_tensor.shape), "shape": repr(value._local_tensor.shape),
"dtype": repr(value._local_tensor.dtype), "dtype": repr(value._local_tensor.dtype),
"value": _sanitize_repr_for_diff(repr(value)), "value": val_repr,
} }
if value._local_tensor.dtype in {torch.float16, torch.float32, torch.bfloat16}:
value = value._local_tensor.clone()
out.update(
{
"mean": _sanitize_repr_for_diff(repr(value.mean())),
"std": _sanitize_repr_for_diff(repr(value.std())),
"min": _sanitize_repr_for_diff(repr(value.min())),
"max": _sanitize_repr_for_diff(repr(value.max())),
}
)
return out
if isinstance(value, torch.Tensor): if isinstance(value, torch.Tensor):
# standard PyTorch Tensor torch.set_printoptions(sci_mode=True)
# return also the shape of such val_repr = _repr_to_list(value)
return {"shape": repr(value.shape), "dtype": repr(value.dtype), "value": _sanitize_repr_for_diff(repr(value))} out = {
"shape": repr(value.shape),
"dtype": repr(value.dtype),
"value": val_repr,
}
if value.dtype in {torch.float16, torch.float32, torch.bfloat16}:
out.update(
{
"mean": _sanitize_repr_for_diff(repr(value.mean())),
"std": _sanitize_repr_for_diff(repr(value.std())),
"min": _sanitize_repr_for_diff(repr(value.min())),
"max": _sanitize_repr_for_diff(repr(value.max())),
}
)
return out
# fallback for everything else (bool, int, float, None, or custom class)
return _sanitize_repr_for_diff(repr(value)) return _sanitize_repr_for_diff(repr(value))
def _repr_to_list(value: torch.Tensor):
"""
Converts a tensor into a sanitized multi-line string representation.
Args:
value (`torch.Tensor`): The tensor to represent.
Returns:
`List[str]`: List of string lines representing the tensor.
"""
torch.set_printoptions(sci_mode=True, linewidth=120)
with StringIO() as buf, redirect_stdout(buf):
print(value) # to redirected stdout to avoid line splits
raw = buf.getvalue()
return _sanitize_repr_for_diff(raw).splitlines()
def prune_outputs_if_children(node): def prune_outputs_if_children(node):
# if there are children, remove this node's "outputs" # if there are children, remove this node's "outputs"
# so we only see outputs at the leaf level # so we only see outputs at the leaf level
@@ -111,22 +153,106 @@ def prune_outputs_if_children(node):
prune_outputs_if_children(child) prune_outputs_if_children(child)
LAYER_SUFFIX_RE = re.compile(r"(.*)\.(\d+)$") # should be generic enough, ends with a number
def is_layer_block(node):
"""
Checks whether a node represents a layer block with submodules.
Args:
node (`dict`): A node from the call tree.
Returns:
`bool`: Whether the node is a layer block.
"""
match = LAYER_SUFFIX_RE.match(node.get("module_path", ""))
if not match or not node.get("children"):
return False
number = match.group(2)
return any(f".{number}." in child.get("module_path", "") for child in node["children"])
def prune_intermediate_layers(node):
"""
Recursively removes intermediate layers from the tree to improve readability.
Keeps at least the first and last layers if many consecutive layers are present.
Args:
node (`dict`): The root or subnode to prune recursively.
"""
if not node.get("children"):
return
layer_blocks = [(i, child) for i, child in enumerate(node["children"]) if is_layer_block(child)]
if len(layer_blocks) > 2:
to_remove = [i for i, _ in layer_blocks[1:-1]]
node["children"] = [child for i, child in enumerate(node["children"]) if i not in to_remove]
for child in node["children"]:
prune_intermediate_layers(child)
def log_model_debug_trace(debug_path, model): def log_model_debug_trace(debug_path, model):
if debug_path: if debug_path:
try: try:
os.makedirs(debug_path, exist_ok=False) os.makedirs(debug_path, exist_ok=True)
output_path = os.path.join(debug_path, model._debugger_module_dump_name + "_debug_tree.json") base = os.path.join(debug_path, model._debugger_module_dump_name + "_debug_tree")
except Exception as e: except Exception as e:
raise ValueError(f"Unexpected or existing debug_path={debug_path}. {e}") raise ValueError(f"Unexpected or existing debug_path={debug_path}. {e}")
else: else:
output_path = model._debugger_module_dump_name + "_debug_tree.json" base = model._debugger_module_dump_name + "_debug_tree"
logger.info(f"Writing model trace at {output_path}")
with open(output_path, "w") as outfile: logger.info(f"Writing model trace at {base}.json")
full_path = base + "_FULL_TENSORS.json"
summary_path = base + "_SUMMARY.json"
prune_outputs_if_children(model._call_tree) prune_outputs_if_children(model._call_tree)
json.dump(model._call_tree, outfile, indent=2)
with open(full_path, "w") as f:
json.dump(model._call_tree, f, indent=2)
# summary-only version for readability - traversing the tree again #TODO optimize?
def strip_values(node):
def clean(val):
if isinstance(val, dict):
val.pop("value", None)
for v in val.values():
clean(v)
elif isinstance(val, list):
for item in val:
clean(item)
clean(node.get("inputs", {}))
clean(node.get("outputs", {}))
for child in node.get("children", []):
strip_values(child)
tree_copy = json.loads(json.dumps(model._call_tree)) # deep copy
strip_values(tree_copy)
with open(summary_path, "w") as f:
json.dump(tree_copy, f, indent=2)
def _attach_debugger_logic(model, class_name, debug_path: str): def _attach_debugger_logic(
model,
debug_path: Optional[str] = ".",
do_prune_layers: Optional[bool] = True,
):
"""
Attaches a debugging wrapper to every module in the model.
This records structured inputs and outputs during the forward pass into a call tree.
Args:
model (`PreTrainedModel`, `nn.Module`): Model to wrap.
debug_path (`str`): Optional directory to dump debug JSON files.
do_prune_layers (`bool`, *optional*, defaults to `True`): Whether to prune intermediate layers.
"""
class_name = model.__class__.__name__
# Prepare data structures on the model object # Prepare data structures on the model object
model._call_tree = {"module_path": class_name, "inputs": None, "outputs": None, "children": []} model._call_tree = {"module_path": class_name, "inputs": None, "outputs": None, "children": []}
model._debugger_model_call_stack = [] model._debugger_model_call_stack = []
@@ -147,7 +273,7 @@ def _attach_debugger_logic(model, class_name, debug_path: str):
"children": [], "children": [],
} }
model._debugger_model_call_stack.append(node) model._debugger_model_call_stack.append(node)
with torch.inference_mode(): with torch.no_grad():
out = orig_forward(*inps, **kws) out = orig_forward(*inps, **kws)
if _is_rank_zero(): if _is_rank_zero():
@@ -188,7 +314,6 @@ def _attach_debugger_logic(model, class_name, debug_path: str):
model._debugger_model_call_stack.append(top_node) model._debugger_model_call_stack.append(top_node)
out = real_top_forward(*inps, **kws) out = real_top_forward(*inps, **kws)
if _is_rank_zero() and model._debugger_model_call_stack: if _is_rank_zero() and model._debugger_model_call_stack:
top_node["outputs"] = _serialize_io(out) top_node["outputs"] = _serialize_io(out)
finished = model._debugger_model_call_stack.pop() finished = model._debugger_model_call_stack.pop()
@@ -198,98 +323,24 @@ def _attach_debugger_logic(model, class_name, debug_path: str):
# prune empty stuff for visibility # prune empty stuff for visibility
[model._call_tree.pop(k, None) for k in list(model._call_tree.keys()) if not model._call_tree[k]] [model._call_tree.pop(k, None) for k in list(model._call_tree.keys()) if not model._call_tree[k]]
# prune layers that are not 0 or last
if do_prune_layers:
prune_intermediate_layers(model._call_tree)
# Write final JSON trace here
log_model_debug_trace(debug_path=debug_path, model=model)
return out return out
model.forward = top_wrapped_forward model.forward = top_wrapped_forward
# Final hook for writing JSON on forward-end
def final_hook(_, inputs, outputs):
if _is_rank_zero() and model._debugger_model_call_stack:
finished = model._debugger_model_call_stack.pop()
model._call_tree["inputs"] = finished["inputs"]
model._call_tree["outputs"] = finished["outputs"]
model._call_tree["children"] = finished["children"]
if _is_rank_zero():
log_model_debug_trace(debug_path=debug_path, model=model)
model.register_forward_hook(final_hook)
# Optionally also for a couple possible hooks that have specific names. It should be just one.
# This means modules that are not typically called "forward" within the model. But we should not need to recurse
# through them.
possible_model_calls = ["language_model", "model"]
for model_call in possible_model_calls:
this_model_call = getattr(model, model_call, None)
if this_model_call and isinstance(this_model_call, (nn.Module, PreTrainedModel)):
this_model_call.register_forward_hook(final_hook)
break # exit the loop after finding one (unsure, but should be just one call.)
@requires(backends=("torch",))
def model_addition_debugger(cls):
"""
# Model addition debugger - a model adder tracer
This decorator is a power user tool intended for model adders.
It tracks all forward calls within a model forward and logs a slice of each input and output on a nested Json.
To note, this decorator enforces `torch.inference_mode()`.
## Usage
add decorator to your model class
```python
from ...modeling_utils import model_addition_debugger
@model_addition_debugger
class MyModel(nn.Module) # Can inherit from PreTrainedModel too
# ... nothing else changes
```
Then, in a separate script (example is for Llava)
```python
import torch
from PIL import Image
import requests
from transformers import LlavaProcessor, LlavaForConditionalGeneration
torch.random.manual_seed(673)
# load pretrained model and processor
model_id = "llava-hf/llava-1.5-7b-hf"
processor = LlavaProcessor.from_pretrained(model_id)
model = LlavaForConditionalGeneration.from_pretrained(model_id, low_cpu_mem_usage=True)
# create random image input
random_image = Image.fromarray(torch.randint(0, 256, (224, 224, 3), dtype=torch.uint8).numpy())
# prompt
prompt = "<image>Describe this image."
# process inputs
inputs = processor(text=prompt, images=random_image, return_tensors="pt")
# call forward method (not .generate!)
with torch.no_grad():
output = model.forward(**inputs)
```
"""
orig_init = cls.__init__
@functools.wraps(cls.__init__)
def wrapped_init(self, *args, **kwargs):
orig_init(self, *args, **kwargs)
_attach_debugger_logic(self, cls.__name__)
cls.__init__ = wrapped_init
return cls
@requires(backends=("torch",)) @requires(backends=("torch",))
@contextmanager @contextmanager
def model_addition_debugger_context(model, debug_path: Optional[str] = None): def model_addition_debugger_context(model, debug_path: Optional[str] = None, do_prune_layers: Optional[bool] = True):
""" """
# Model addition debugger - context manager for model adders # Model addition debugger - context manager for model adders
This context manager is a power user tool intended for model adders. This context manager is a power user tool intended for model adders.
It tracks all forward calls within a model forward and logs a slice of each input and output on a nested Json. It tracks all forward calls within a model forward and logs a slice of each input and output on a nested Json.
To note, this context manager enforces `torch.inference_mode()`. To note, this context manager enforces `torch.no_grad()`.
## Usage ## Usage
@@ -300,6 +351,7 @@ def model_addition_debugger_context(model, debug_path: Optional[str] = None):
from PIL import Image from PIL import Image
import requests import requests
from transformers import LlavaProcessor, LlavaForConditionalGeneration from transformers import LlavaProcessor, LlavaForConditionalGeneration
from transformers.model_debugging_utils import model_addition_debugger_context
torch.random.manual_seed(673) torch.random.manual_seed(673)
# load pretrained model and processor # load pretrained model and processor
@@ -317,13 +369,16 @@ def model_addition_debugger_context(model, debug_path: Optional[str] = None):
inputs = processor(text=prompt, images=random_image, return_tensors="pt") inputs = processor(text=prompt, images=random_image, return_tensors="pt")
# call forward method (not .generate!) # call forward method (not .generate!)
with model_addition_debugger_context(model): with model_addition_debugger_context(model, debug_path="Your_debug_path", do_prune_layers=False):
output = model.forward(**inputs) output = model.forward(**inputs)
``` ```
""" """
_attach_debugger_logic(model, model.__class__.__name__, debug_path) orig_forwards = {m: m.forward for _, m in model.named_modules()}
orig_forwards[model] = model.forward
_attach_debugger_logic(model, debug_path, do_prune_layers)
try: try:
yield model yield model
finally: finally:
pass for module_instance, forward_method in orig_forwards.items():
module_instance.forward = forward_method

4
src/transformers/utils/dummy_pt_objects.py
View File

@@ -538,10 +538,6 @@ def convert_and_export_with_cache(*args, **kwargs):
requires_backends(convert_and_export_with_cache, ["torch"]) requires_backends(convert_and_export_with_cache, ["torch"])
def model_addition_debugger(*args, **kwargs):
requires_backends(model_addition_debugger, ["torch"])
def model_addition_debugger_context(*args, **kwargs): def model_addition_debugger_context(*args, **kwargs):
requires_backends(model_addition_debugger_context, ["torch"]) requires_backends(model_addition_debugger_context, ["torch"])

122
tests/utils/test_model_debugging_utils.py Normal file
View File

@@ -0,0 +1,122 @@
# coding=utf-8
# Copyright 2025 The HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import gc
import json
import os
import tempfile
import unittest
from pathlib import Path
from transformers import is_torch_available
from transformers.model_debugging_utils import model_addition_debugger_context
if is_torch_available():
import torch
from torch import nn
class ToyModel(nn.Module):
def __init__(self):
super().__init__()
self.embed = nn.Embedding(10, 4)
self.linear_1 = nn.Linear(4, 8)
self.linear_2 = nn.Linear(8, 2)
self.act = nn.ReLU()
def forward(self, input_ids: str):
hidden_states = self.embed(input_ids).mean(dim=1)
hidden_states = self.act(self.linear_1(hidden_states))
return self.linear_2(hidden_states)
class TestModelAdditionDebugger(unittest.TestCase):
def setUp(self):
self.model = ToyModel()
self.inputs = {"input_ids": torch.randint(0, 10, (1, 3))}
def tearDown(self):
gc.collect()
def test_debugger_outputs(self):
with tempfile.TemporaryDirectory() as tmpdir:
with model_addition_debugger_context(self.model, debug_path=str(tmpdir)):
_ = self.model.forward(**self.inputs)
base = f"{self.model.__class__.__name__}_debug_tree"
summary = Path(os.path.join(tmpdir, f"{base}_SUMMARY.json"))
full = Path(os.path.join(tmpdir, f"{base}_FULL_TENSORS.json"))
self.assertTrue(os.path.isfile(summary) and os.path.isfile(full))
data = json.loads(summary.read_text())
self.assertTrue({"module_path", "inputs", "children"} <= data.keys())
self.assertTrue(data["children"])
class ToyLayer(nn.Module):
def __init__(self, layer_index):
super().__init__()
self.layer_index = layer_index
self.layer_operation = nn.Linear(4, 4)
def forward(self, hidden_states):
return self.layer_operation(hidden_states)
class ToyModelWithLayers(nn.Module):
def __init__(self):
super().__init__()
self.input_proj = nn.Linear(4, 4)
self.layers = nn.ModuleList([ToyLayer(layer_index) for layer_index in range(6)])
self.output_proj = nn.Linear(4, 2)
def forward(self, x):
x = self.input_proj(x)
for layer in self.layers:
x = layer(x)
return self.output_proj(x)
class TestModelWithLayers(unittest.TestCase):
def setUp(self):
self.inputs = {"input_ids": torch.randint(0, 10, (1, 3))}
self.model_with_layers = ToyModelWithLayers()
self.dense_input = {"x": torch.randn(1, 4)}
def tearDown(self):
gc.collect()
def test_layer_pruning_behavior(self):
# No pruning: expect all 6 layers
with tempfile.TemporaryDirectory() as tmpdir:
with model_addition_debugger_context(self.model_with_layers, debug_path=tmpdir, do_prune_layers=False):
_ = self.model_with_layers(**self.dense_input)
summary_path = os.path.join(tmpdir, "ToyModelWithLayers_debug_tree_SUMMARY.json")
with open(summary_path) as f:
data = json.load(f)
self.assertEqual(set(data.keys()), {"module_path", "inputs", "children"})
for layer_index in range(6):
self.assertEqual(
data["children"][layer_index + 1]["module_path"],
f"ToyModelWithLayers.layers.{int(layer_index)}",
)
# Pruning: expect only 2 layers (0 and 5)
with tempfile.TemporaryDirectory() as tmpdir:
with model_addition_debugger_context(self.model_with_layers, debug_path=tmpdir, do_prune_layers=True):
_ = self.model_with_layers(**self.dense_input)
summary_path = os.path.join(tmpdir, "ToyModelWithLayers_debug_tree_SUMMARY.json")
with open(summary_path) as f:
data = json.load(f)
self.assertEqual(set(data.keys()), {"module_path", "inputs", "children"})
self.assertEqual(data["children"][1]["module_path"], "ToyModelWithLayers.layers.0")
self.assertEqual(data["children"][2]["module_path"], "ToyModelWithLayers.layers.5")