Comments (2)
I also tried with the load function in vit.
from big_vision.models.vit import load as vit_load
init_params = vit_load(None, 'clippo_b16_yfcc100m_i21k_init_75c4.npz', None)
and in vit_load
def fix_old_checkpoints(params):
"""Fix small bwd incompat that can't be resolved with names in model def."""
params = flax.core.unfreeze(flax.training.checkpoints.convert_pre_linen(params))
# Original ViT paper variant had posemb in a module:
if "posembed_input" in params["img"]: # IT WAS params['Transformer'] before which I replaced as no transformer key in weights
logging.info("ViT: Loading and fixing VERY old posemb")
posemb = params["img"].pop("posembed_input")
params["pos_embedding"] = posemb["pos_embedding"]
# Widely used version before 2022 had posemb in Encoder:
if "pos_embedding" in params["img"]:
logging.info("ViT: Loading and fixing old posemb")
params["pos_embedding"] = params["img"].pop("pos_embedding")
# Old vit.py used to first concat [cls] token, then add posemb.
# This means a B/32@224px would have 7x7+1 posembs. This is useless and clumsy
# so we changed to add posemb then concat [cls]. We can recover the old
# checkpoint by manually summing [cls] token and its posemb entry.
if "pos_embedding" in params:
pe = params["pos_embedding"]
if int(np.sqrt(pe.shape[1])) ** 2 + 1 == int(pe.shape[1]):
logging.info("ViT: Loading and fixing combined cls+posemb")
pe_cls, params["pos_embedding"] = pe[:, :1], pe[:, 1:]
if "cls" in params:
params["cls"] += pe_cls
# MAP-head variants during ViT-G development had it inlined:
if "probe" in params:
params["MAPHead_0"] = {
k: params.pop(k)
for k in [
"probe",
"MlpBlock_0",
"MultiHeadDotProductAttention_0",
"LayerNorm_0",
]
}
return params
def load(
init_params, init_file, model_cfg, dont_load=()
): # pylint: disable=invalid-name because we had to CamelCase above.
"""Load init from checkpoint, both old model and this one. +Hi-res posemb."""
del model_cfg
init_file = VANITY_NAMES.get(init_file, init_file)
restored_params = utils.load_params(None, init_file)
restored_params = fix_old_checkpoints(restored_params)
# possibly use the random init for some of the params (such as, the head).
restored_params = common.merge_params(restored_params, init_params, dont_load)
# resample posemb if needed.
if init_params and "pos_embedding" in init_params:
restored_params["pos_embedding"] = resample_posemb(
old=restored_params["pos_embedding"], new=init_params["pos_embedding"]
)
return restored_params
I still get the same error:
---------------------------------------------------------------------------
ScopeParamNotFoundError Traceback (most recent call last)
Cell In[10], [line 48](vscode-notebook-cell:?execution_count=10&line=48)
[45](vscode-notebook-cell:?execution_count=10&line=45) labels = batch['label'].numpy()
[47](vscode-notebook-cell:?execution_count=10&line=47) # Forward pass
---> [48](vscode-notebook-cell:?execution_count=10&line=48) zimg, _ = model.apply({'params': init_params}, images)
[50](vscode-notebook-cell:?execution_count=10&line=50) # Get logits
[51](vscode-notebook-cell:?execution_count=10&line=51) logits = classifier.apply({'params': classifier.params}, zimg)
[... skipping hidden 6 frame]
File [~/Desktop/projects/big_vision/big_vision/models/vit.py:186](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:186), in _Model.__call__(self, image, train)
[183](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:183) out = {}
[185](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:185) # Patch extraction
--> [186](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:186) x = out["stem"] = nn.Conv(
[187](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:187) self.width,
[188](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:188) self.patch_size,
[189](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:189) strides=self.patch_size,
[190](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:190) padding="VALID",
[191](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:191) name="embedding",
[192](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:192) )(image)
[194](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:194) n, h, w, c = x.shape
[195](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/Desktop/projects/big_vision/big_vision/models/vit.py:195) x = jnp.reshape(x, [n, h * w, c])
[... skipping hidden 2 frame]
File [~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:480](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:480), in _Conv.__call__(self, inputs)
[474](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:474) if self.mask is not None and self.mask.shape != kernel_shape:
[475](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:475) raise ValueError(
[476](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:476) 'Mask needs to have the same shape as weights. '
[477](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:477) f'Shapes are: {self.mask.shape}, {kernel_shape}'
[478](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:478) )
--> [480](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:480) kernel = self.param(
[481](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:481) 'kernel', self.kernel_init, kernel_shape, self.param_dtype
[482](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:482) )
[484](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:484) if self.mask is not None:
[485](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/linen/linear.py:485) kernel *= self.mask
[... skipping hidden 1 frame]
File [~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:896](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:896), in Scope.param(self, name, init_fn, unbox, *init_args)
[894](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:894) if self.is_collection_empty('params'):
[895](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:895) raise errors.ScopeCollectionNotFound('params', name, self.path_text)
--> [896](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:896) raise errors.ScopeParamNotFoundError(name, self.path_text)
[897](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:897) value = init_fn(self.make_rng('params'), *init_args)
[898](https://file+.vscode-resource.vscode-cdn.net/home/ahmad/Desktop/projects/big_vision/big_vision/trainers/~/anaconda3/envs/robustness/lib/python3.8/site-packages/flax/core/scope.py:898) self.put_variable('params', name, value)
ScopeParamNotFoundError: Could not find parameter named "kernel" in scope "/embedding". (https://flax.readthedocs.io/en/latest/api_reference/flax.errors.html#flax.errors.ScopeParamNotFoundError)
How ever, there is now an extra key(pos_embedding) in init_params with this dict_keys(['img', 't', 'pos_embedding'])
as we created it in the function
from big_vision.
Hi,
I think the issue you are observing is because CLIPPO uses a wrapper for models.vit
(to do two forward passes with natural and text images), namely models.proj.clippo.one_tower
. Loading the checkpoints into this model should work.
You can follow the steps in the colab. This might be broken at head, but using your method to check out the older commit (or the follow up commit which adds the colab) should work.
from big_vision.
Related Issues (20)
- Negative rho values in GSAM training HOT 1
- FlexiVit is also flexible with image resolution?
- Mixup Per Example?
- Clarification: SigLIP Image Transform
- Errors in notebooks
- Confusion on FlexiViT
- Error with putting arrays on CPU in cloud TPUs HOT 1
- How to save fine tuned PaliGemma model? HOT 3
- [QUESTION] How to perform inference on trained model? HOT 8
- Can I convert paligemma npz model to pytorch to safetensors? HOT 3
- Implementation of contrast() seems wrong
- Behavior of `solarize()` depends on integer overflow
- No ChartQA ft weights released on Kaggle HOT 1
- PlaiGemma finetuned model HOT 1
- Loss Scale for Training Siglip
- [Question] How to inference / captioning short video? HOT 4
- Question about the evaluation on instance segmentation about PaliGemma
- Could you please give me an instruction on how to use paligemma's evaluator?
- What's a "`rezied` method"?
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from big_vision.