jinaai/jina-clip-v2 · Discussions

Encode image gives nan with transformers > 5

#63 opened 3 months ago by

Ps06600

CoreMLExecutionProvider fails with jina-clip-v2 ONNX on Apple Silicon

#62 opened 3 months ago by

ghs260

encode_image returns embeddings with NaN values in Transformers 5

#61 opened 5 months ago by

yelboudouri

Dependencies while exporting into ONNX format.

#59 opened 8 months ago by

jayn2u

Using task (LoRA) with ONNX model?

#58 opened 11 months ago by

satorean

Use text and image encoder separately with onnxruntime

6

#57 opened about 1 year ago by

Frayin

Pointer argument (at 0) cannot be accessed from Triton (cpu tensor?)

#56 opened over 1 year ago by

vivian345

Blocking access to onnx backend?

#55 opened over 1 year ago by

rcaulk

'JinaCLIPConfig' object has no attribute 'num_attention_heads'

1

#54 opened over 1 year ago by

SupercarryNg

Inconsistent image processing times

4

#53 opened over 1 year ago by

Peniak

When running the sample locally, an error was reported：AttributeError: 'EVAVisionTransformer' object has no attribute '_initialize_weights'

5

#52 opened over 1 year ago by

yslovesby

how to load the jina-clip-v2 model offline? help me plz!!

2

#51 opened over 1 year ago by

HT-NEKO

Failed to load the downloaded model locally

6

#50 opened over 1 year ago by

HT-NEKO

Encountering issues while deploying Jina-CLIP v2 on AWS using jina-sagemaker.

4

#49 opened over 1 year ago by

YanaKnudsen

Will you release the fine-tuning code as open source?

1

#48 opened over 1 year ago by

classdemo

gguf format please?

👍 1

1

#46 opened over 1 year ago by

kalle07

Using a slow image processor as `use_fast` is unset - Transformers issue

🚀❤️ 3

#45 opened over 1 year ago by

do-me

ValueError: Invalid image type. Expected either PIL.Image.Image, numpy.ndarray, torch.Tensor, tf.Tensor or jax.ndarray, but got <class 'list'>.

#44 opened over 1 year ago by

do-me

Running on Apple Silicon with GPU

1

#43 opened over 1 year ago by

do-me

do you have 224x224 or 386x386 pretrained clip model?

2

#42 opened over 1 year ago by

bojohn

Ollama use?

1

#41 opened over 1 year ago by

YarvixPA

Does “multilingual MSCOCO” refer to “XTD”?

1

#38 opened over 1 year ago by

Chao985

Onnx clip-v2 inferrence q4,uint8 speed slower than fp16

#37 opened over 1 year ago by

eminarcissus

Loading jina-clip-v2 extremely slow

1

#36 opened over 1 year ago by

kenshinn

The issue of VRAM usage in image extraction

1

#35 opened over 1 year ago by

jimmyseven

Jina Clip V2 Encode_image latency is proportional to batch size

3

#33 opened over 1 year ago by

lssatvik

Jina Clip V2: Inconsistent Embeddings

14

#32 opened over 1 year ago by

lssatvik

Dose this model support mixed-modality input?

1

#29 opened over 1 year ago by

Haon-Chen

ValueError: JinaCLIPModel does not support Flash Attention 2.0 yet

1

#27 opened over 1 year ago by

kelvennn

Exception: data did not match any variant of untagged enum PyPreTokenizerTypeWrapper at line 83 column 3

1

#25 opened over 1 year ago by

h3clikejava

Transformers.js error: can't find model.onnx_data file

4

#22 opened almost 2 years ago by

recallapp

ValueError: class 'jina_clip_v2.configuration_clip.JinaCLIPConfig'>

1

#21 opened almost 2 years ago by

zhoudl5

onnx use pixel_values = np.random.rand(0, 3, 512, 512).astype(np.float32) will cause onnx warning

4

#15 opened almost 2 years ago by

xxai-art