This PR batch the fix for a few accuracy failures issues during training by raising tolerance. I do that only for models that I think it fails not due to real issue.
## sebotnet33ts_256
The accuracy test for this model start to fail around June 05 [link](https://hud.pytorch.org/benchmark/timm_models/inductor_with_cudagraphs?dashboard=torchinductor&startTime=Sun%2C%2002%20Jun%202024%2007%3A19%3A38%20GMT&stopTime=Tue%2C%2002%20Jul%202024%2007%3A19%3A38%20GMT&granularity=day&mode=training&dtype=amp&lBranch=main&lCommit=04a0d856207d83c2031e4b9cb6825ba3e0092850&rBranch=main&rCommit=e62925930f6a62f6aeeb1fe1a661a9bd3352b53d&model=sebotnet33ts_256).
I can not repro locally, but from the log from the dashboard:
```
RMSE (res-fp64): 0.09441, (ref-fp64): 0.02971 and shape=torch.Size([1536]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.040000
```
raising the tolerance should fix it.
## DebertaForQuestionAnswering
This model fails accuracy test on the dashboard only in max-autotune mode. I can not repro locally by command:
```
TORCHINDUCTOR_MAX_AUTOTUNE=1 time python benchmarks/dynamo/huggingface.py --accuracy --no-translation-validation --training --amp --backend inductor --device cuda --only DebertaForQuestionAnswering
```
From error message on the dashboard:
```
RMSE (res-fp64): 0.01803, (ref-fp64): 0.00537 and shape=torch.Size([2]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.010000
```
0.02 tolerance should suppress this error.
## gluon_inception_v3
This model fail on the dashboard in max-autotune mode. I can not repro locally by command
```
TORCHINDUCTOR_MAX_AUTOTUNE=1 time python benchmarks/dynamo/timm_models.py --accuracy --training --amp --backend inductor --disable-cudagraphs --device cuda --only gluon_inception_v3
```
From error message on the dashboard
```
RMSE (res-fp64): 0.02798, (ref-fp64): 0.00730 and shape=torch.Size([384]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.010000
Accuracy failed for key name Mixed_7c.branch3x3dbl_3a.bn.running_var
```
raising tolerance should suppress this error.
# mobilenetv3_large_100
Fail in MA model. I can not repro locally by command
```
TORCHINDUCTOR_MAX_AUTOTUNE=1 time python benchmarks/dynamo/timm_models.py --accuracy --training --amp --backend inductor --disable-cudagraphs --device cuda --only
```
The error message on the dashboard is
```
RMSE (res-fp64): 0.29754, (ref-fp64): 0.05205 and shape=torch.Size([]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.040000
```
The tensor is so small that the noise can be high. I use larger multiplier for smaller tensor in torch._dynamo.utils.same.
# yolov3
Fail on dashboard with error
```
Error on the dashboard: RMSE (res-fp64): 0.01278, (ref-fp64): 0.00246 and shape=torch.Size([256]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.001000
```
Fix it by using a larger multiplier for smaller tensors and raising the tolereance.
# timm_efficientdet
Fail on the dashboard with error
```
E0623 18:37:43.638000 139924418725056 torch/_dynamo/utils.py:1468] RMSE (res-fp64): 0.00096, (ref-fp64): 0.00009 and shape=torch.Size([2]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.001000
```
But I can not repro locally with command
```
time python benchmarks/dynamo/torchbench.py --backend inductor --amp --performance --only timm_efficientdet --training
```
Raise the tolerance should fix.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129941
Approved by: https://github.com/jansel
ghstack dependencies: #129996
I'm debugging the accuracy failure for training vision_maskrcnn.
Unfortunately I could not succeed to run it locally (I've check pined commits for torchbenchmars/torchvision are correct, and reinstalled torchbenchmark for mask_rcnn). I get this error:
```
eager run fail: AssertionError: targets should not be none when in training mode
```
(Command: time python benchmarks/dynamo/torchbench.py --backend inductor --amp --performance --training --only vision_maskrcnn )
But look at the log from the dashboard
```
E0623 19:17:59.085000 140114670171328 torch/_dynamo/utils.py:1468] RMSE (res-fp64): nan, (ref-fp64): nan and shape=torch.Size([1024, 256, 1, 1]). res.dtype: torch.float32, multiplier: 3.000000, tol: 0.001000
```
We can see both the reference number and the pt2 number are NaN. I change torch._dynamo.utils.same to return true if both RMSE values are NaN.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129996
Approved by: https://github.com/jansel
# Compile time for eager backend
## AlbertForMaskedLM
No inlining - 3.65 seconds
Inlining on main - 7.48 seconds
Inlining + this PR - 6.70 seconds
## MobileBertForMaskedLM
No inlining - 26.90 seconds
Inlining on main - 48.21 seconds
Inlining + this PR - 43.85 seconds
*Next PR in the stack makes the total compile time better/comparable to no inlining*
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129315
Approved by: https://github.com/jansel
ghstack dependencies: #129316
FIXES#113263. Same idea as in https://github.com/pytorch/pytorch/pull/113417, but we need a more intrusive C API to silently nop default saved tensor hooks, in order to support user-code that use torch.autograd.disable_saved_tensors_hooks (see test_unpack_hooks_can_be_disabled). We mock the output of get_hooks while leaving push/pop untouched.
For compiled autograd, we're firing pack hooks once and unpack hooks twice right now, I'll look into this separately from this issue.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/123196
Approved by: https://github.com/soulitzer
This is a short-term fix (for 2.4). In the longer term we should
fix https://github.com/pytorch/pytorch/issues/128430
The problem is that warnings.warn that are inside Dynamo print
all the time. Python warnings are supposed to print once, unless their
cache is reset: Dynamo ends up resetting that cache everytime it runs.
As a workaround we provide our own warn_once cache that is keyed on the
warning msg. I am not worried about this increasing memory usage because
that's effectively what python's warnings.warn cache does.
Test Plan:
- fix tests.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/128456
Approved by: https://github.com/anijain2305
Tracing through `__init__` is important because it initializes (calls STORE_ATTR) on members. By doing that, we kick in the mutation tracking for these objects. So, things like mutating `_modules` etc is tracked automatically.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/126578
Approved by: https://github.com/jansel
ghstack dependencies: #128001
To fix data-dependent errors we want to recommend that people use `torch._check*` APIs. The `constrain_as*` APIs should be fully subsumed by them, and in the future we should kill them entirely.
Differential Revision: D56774333
Pull Request resolved: https://github.com/pytorch/pytorch/pull/125253
Approved by: https://github.com/ezyang
While studying some tlparse, I noticed that CompilationMetrics was reporting that there was no error for frames that have no nodes. I'm pretty sure we don't actually install a frame in this situation. has_guarded_code will tell us if that's the case, because it says if the GuardedCode object is None or not.
Actually, while working on this, I was wondering if we can ever trigger the "skip this frame entirely, do not trace it ever again" codepath, as best as I could tell, it's impossible for this to happen by the time we get to compilation metrics block.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/125279
Approved by: https://github.com/yanboliang
The important comment:
```
# Whenever we allocate a fresh unbacked Symbol, we add it to this
# pending list. Unbacked symbol allocation can occur at unpredictable
# points during meta tensor propagation, but at some point, the we
# have to know what the binding site for an unbacked symbol is, and
# this is computed when we actually place the node in the graph. The
# important thing is that we always actually handle every unaccounted
# for unbacked symbol, so this list helps us keep track of them and
# then make sure they are all accounted for.
#
# We could potentially give rise to errors earlier by lexically
# scoping when we do propagation, and only allowing unbacked symbols
# to be allocated at this point in time. However this is inconvenient
# to do in Dynamo, because fake tensor propagation is far from when we
# analyze binding sites (set_example_value), so we do it in a more
# mutatey way.
#
# NB: fresh unbacked symbols NEVER get substitutions applied to them,
# they are binding sites!
```
The compute_unbacked_bindings is the other half of the equation: the thing that actually consumes the pending_fresh_unbacked_symbols and does something with them. Important comment:
```
After having run fake tensor propagation and producing example_value
result, traverse example_value looking for freshly bound unbacked
symbols and record their paths for later. It is an error if
we have allocated an unbacked SymInt but it cannot be found in
example_value. (NB: this means if you have a multi-output
function, you must call this on the tuple of tensor output, you
cannot wait!)
```
For example, if I return a tensor with size `[u0, u1]`, and u1 is a fresh unbacked SymInt, then I'll have `{u1: KeyPath(".size(1)")}`, telling me I can get u1 by running `size(1)` on the result of this node. u0 is not fresh (it probably flowed in as an argument), so I don't generate a binding for it.
I eventually intend to propagate this information all the way to Inductor lowering, where extra metadata about unbacked symbol binding will be canonically used for codegen, instead of trying to infer it from defs/uses.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/124290
Approved by: https://github.com/lezcano
I'm going to setup some extra behavior when we set example value, so
I need a convenient place to interpose. I cannot easily do it on
meta itself because its a generic dict with no interposition point.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/124176
Approved by: https://github.com/oulgen
ghstack dependencies: #124105, #124059
Automatic fixes that replaces certain list comprehensions with generator ones where appropriate so that they are immediately consumed. This is preview functionality in ruff for rule C419 and it was automatically applied.
Co-authored-by: Nikita Shulga <2453524+malfet@users.noreply.github.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/123960
Approved by: https://github.com/malfet
### Context
In today's Dynamo, we lift all tensors encountered during tracing to be individual graph inputs, even when they were in a container.
And [Dynamo generates](fdc281f258/torch/_dynamo/codegen.py (L371)) the runtime function's signature using the graph's graphargs.
This means that the generated function will have each grapharg as an argument, which is problematic if we want to free the inputs in inductor codegen. See [python function arguments are kept alive for the duration of the function call](https://github.com/pytorch/pytorch/pull/83137#issuecomment-1211320670).
```python
# original code
def forward(inputs):
a, b, c, d, e = inputs
inputs.clear()
out = a
out += b
del b # frees memory
out += c
del c # frees memory
out += d
del d # frees memory
out += e
del e # frees memory
return out
# compiled code:
def forward(a, b, c, d, e):
# b, c, d, e can't be freed before end of function
```
This isn't a concern when compiling forward because a, b, c, d, e are all from user code, and should be kept alive. But when compiling backwards, a, b, c, d, e may be intermediate results i.e. activations, that we DO want to clear ASAP to remain on par with eager peak memory.
### Solution
We have encountered similar memory problems in AOTAutograd before, where we adopted the boxed calling convention (wrapping to-be-freed objects in a list), adding list clearing to inductor codegen, and being careful about holding references to elements in the input list. We need to do something similar, but for inputs from the user program (compiled autograd fx graph in this case).
This PR support lists as graphargs/placeholder nodes. When tracing a list of tensors, we create a node for it, and pre-emptively initialize variable trackers for its elements before they are used in the user program. Subsequent uses of those variables will find hits in the lookup table `input_source_to_var`.
With the inputs as a list in the graph args, our compiled code can free inputs just like in the eager case.
```python
def forward(inputs):
# a, b, c, d, e can be freed within the function now
```
Currently, AOT/Inductor flattens list input via [flatten_graph_inputs wrapper](597f479643/torch/_inductor/compile_fx.py (L1454-L1478)), which is why this PR's CI can be green. Additional changes are needed to its runtime wrapper, done in the next PR. The next step is to ensure that we are careful in forwarding the list to inductor codegen without holding additional references.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/122353
Approved by: https://github.com/jansel
ghstack dependencies: #123630, #123674
`dynamo.explain()` was updated to return a structure but the docs weren't updated to match.
- Update the docs to use the new API
- Remove some dead code left when `explain` was updated.
- Drive-by: Fix some `nopython` uses that I noticed
- Drive-by: I noticed an ignored error coming from CleanupHook on shutdown - make it check the global before setting it.
Fixes#122573
Pull Request resolved: https://github.com/pytorch/pytorch/pull/122745
Approved by: https://github.com/jansel
