Update README.md

2023-03-18 02:23:31 +00:00 · 2023-03-18 02:16:03 +00:00 · 2023-02-27 14:58:18 +00:00 · 2023-02-10 08:36:54 +00:00 · 2023-02-10 08:34:31 +00:00 · 2023-02-09 10:15:07 +00:00
5 changed files with 202 additions and 83 deletions
--- a/README.md
+++ b/README.md
@ -0,0 +1,159 @@
+---
+language: en
+license: mit
+tags:
+- vision
+- image-to-text
+- image-captioning
+- visual-question-answering
+pipeline_tag: image-to-text
+inference: false
+---
+
+# BLIP-2, Flan T5-xl, pre-trained only
+
+BLIP-2 model, leveraging [Flan T5-xl](https://huggingface.co/google/flan-t5-xl) (a large language model).
+It was introduced in the paper [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://arxiv.org/abs/2301.12597) by Li et al. and first released in [this repository](https://github.com/salesforce/LAVIS/tree/main/projects/blip2).
+
+Disclaimer: The team releasing BLIP-2 did not write a model card for this model so this model card has been written by the Hugging Face team.
+
+## Model description
+
+BLIP-2 consists of 3 models: a CLIP-like image encoder, a Querying Transformer (Q-Former) and a large language model.
+
+The authors initialize the weights of the image encoder and large language model from pre-trained checkpoints and keep them frozen
+while training the Querying Transformer, which is a BERT-like Transformer encoder that maps a set of "query tokens" to query embeddings,
+which bridge the gap between the embedding space of the image encoder and the large language model.
+
+The goal for the model is simply to predict the next text token, giving the query embeddings and the previous text.
+
+<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/model_doc/blip2_architecture.jpg"
+alt="drawing" width="600"/> 
+
+This allows the model to be used for tasks like:
+
+- image captioning
+- visual question answering (VQA)
+- chat-like conversations by feeding the image and the previous conversation as prompt to the model
+
+## Direct Use and Downstream Use
+
+You can use the raw model for conditional text generation given an image and optional text. See the [model hub](https://huggingface.co/models?search=Salesforce/blip) to look for
+fine-tuned versions on a task that interests you.
+
+## Bias, Risks, Limitations, and Ethical Considerations
+
+BLIP2-FlanT5 uses off-the-shelf Flan-T5 as the language model. It inherits the same risks and limitations from [Flan-T5](https://arxiv.org/pdf/2210.11416.pdf):
+
+> Language models, including Flan-T5, can potentially be used for language generation in a harmful way, according to Rae et al. (2021). Flan-T5 should not be used directly in any application, without a prior assessment of safety and fairness concerns specific to the application.
+
+BLIP2 is fine-tuned on image-text datasets (e.g. [LAION](https://laion.ai/blog/laion-400-open-dataset/) ) collected from the internet.  As a result the model itself is potentially vulnerable to generating equivalently inappropriate content or replicating inherent biases in the underlying data.
+
+BLIP2 has not been tested in real world applications. It should not be directly deployed in any applications. Researchers should first carefully assess the safety and fairness of the model in relation to the specific context they’re being deployed within.
+
+### How to use
+
+For code examples, we refer to the [documentation](https://huggingface.co/docs/transformers/main/en/model_doc/blip-2#transformers.Blip2ForConditionalGeneration.forward.example).
+
+#### Running the model on CPU
+
+<details>
+<summary> Click to expand </summary>
+
+```python
+import requests
+from PIL import Image
+from transformers import BlipProcessor, Blip2ForConditionalGeneration
+
+processor = BlipProcessor.from_pretrained("Salesforce/blip2-flan-t5-xl")
+model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-flan-t5-xl")
+
+img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg' 
+raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
+
+question = "how many dogs are in the picture?"
+inputs = processor(raw_image, question, return_tensors="pt")
+
+out = model.generate(**inputs)
+print(processor.decode(out[0], skip_special_tokens=True))
+```
+</details>
+
+#### Running the model on GPU
+
+##### In full precision 
+
+<details>
+<summary> Click to expand </summary>
+
+```python
+# pip install accelerate
+import requests
+from PIL import Image
+from transformers import Blip2Processor, Blip2ForConditionalGeneration
+
+processor = Blip2Processor.from_pretrained("Salesforce/blip2-flan-t5-xl")
+model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-flan-t5-xl", device_map="auto")
+
+img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg' 
+raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
+
+question = "how many dogs are in the picture?"
+inputs = processor(raw_image, question, return_tensors="pt").to("cuda")
+
+out = model.generate(**inputs)
+print(processor.decode(out[0], skip_special_tokens=True))
+```
+</details>
+
+##### In half precision (`float16`)
+
+<details>
+<summary> Click to expand </summary>
+
+```python
+# pip install accelerate
+import torch
+import requests
+from PIL import Image
+from transformers import Blip2Processor, Blip2ForConditionalGeneration
+
+processor = Blip2Processor.from_pretrained("Salesforce/blip2-flan-t5-xl")
+model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-flan-t5-xl", torch_dtype=torch.float16, device_map="auto")
+
+img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg' 
+raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
+
+question = "how many dogs are in the picture?"
+inputs = processor(raw_image, question, return_tensors="pt").to("cuda", torch.float16)
+
+out = model.generate(**inputs)
+print(processor.decode(out[0], skip_special_tokens=True))
+```
+</details>
+
+##### In 8-bit precision (`int8`)
+
+<details>
+<summary> Click to expand </summary>
+
+```python
+# pip install accelerate bitsandbytes
+import torch
+import requests
+from PIL import Image
+from transformers import Blip2Processor, Blip2ForConditionalGeneration
+
+processor = Blip2Processor.from_pretrained("Salesforce/blip2-flan-t5-xl")
+model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-flan-t5-xl", load_in_8bit=True, device_map="auto")
+
+img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg' 
+raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
+
+question = "how many dogs are in the picture?"
+inputs = processor(raw_image, question, return_tensors="pt").to("cuda", torch.float16)
+
+out = model.generate(**inputs)
+print(processor.decode(out[0], skip_special_tokens=True))
+```
+</details>
--- a/config.json
+++ b/config.json
@ -94,7 +94,7 @@
    ],
    "bad_words_ids": null,
    "begin_suppress_tokens": null,
-    "bos_token_id": null,
+    "bos_token_id": 1,
    "chunk_size_feed_forward": 0,
    "cross_attention_hidden_size": null,
    "d_ff": 5120,
--- a/pytorch_model-00001-of-00002.bin
+++ b/pytorch_model-00001-of-00002.bin
@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:537bb68668d2cc0e84297ee5c23a77e5a86fc28f528afc642155bf5b02dc7e7f
-size 9441197415
+oid sha256:78da3eebf4732877fd22b73dc2ed57dae30c466d5a44b7c329de376fa4c79a88
+size 9441403325
--- a/pytorch_model.bin.index.json
+++ b/pytorch_model.bin.index.json
@ -1,6 +1,6 @@
 {
  "metadata": {
-    "total_size": 16295951872
+    "total_size": 16296171520
  },
  "weight_map": {
    "language_model.decoder.block.0.layer.0.SelfAttention.k.weight": "pytorch_model-00001-of-00002.bin",
@ -834,9 +834,8 @@
    "vision_model.encoder.layers.0.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.0.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.0.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.0.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.0.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.0.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.0.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.1.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.1.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.1.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -847,9 +846,8 @@
    "vision_model.encoder.layers.1.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.1.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.1.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.1.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.1.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.1.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.1.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.10.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.10.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.10.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -860,9 +858,8 @@
    "vision_model.encoder.layers.10.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.10.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.10.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.10.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.10.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.10.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.10.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.11.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.11.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.11.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -873,9 +870,8 @@
    "vision_model.encoder.layers.11.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.11.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.11.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.11.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.11.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.11.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.11.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.12.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.12.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.12.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -886,9 +882,8 @@
    "vision_model.encoder.layers.12.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.12.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.12.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.12.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.12.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.12.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.12.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.13.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.13.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.13.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -899,9 +894,8 @@
    "vision_model.encoder.layers.13.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.13.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.13.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.13.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.13.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.13.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.13.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.14.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.14.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.14.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -912,9 +906,8 @@
    "vision_model.encoder.layers.14.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.14.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.14.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.14.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.14.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.14.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.14.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.15.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.15.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.15.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -925,9 +918,8 @@
    "vision_model.encoder.layers.15.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.15.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.15.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.15.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.15.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.15.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.15.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.16.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.16.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.16.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -938,9 +930,8 @@
    "vision_model.encoder.layers.16.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.16.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.16.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.16.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.16.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.16.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.16.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.17.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.17.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.17.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -951,9 +942,8 @@
    "vision_model.encoder.layers.17.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.17.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.17.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.17.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.17.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.17.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.17.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.18.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.18.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.18.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -964,9 +954,8 @@
    "vision_model.encoder.layers.18.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.18.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.18.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.18.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.18.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.18.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.18.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.19.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.19.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.19.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -977,9 +966,8 @@
    "vision_model.encoder.layers.19.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.19.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.19.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.19.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.19.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.19.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.19.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.2.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.2.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.2.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -990,9 +978,8 @@
    "vision_model.encoder.layers.2.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.2.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.2.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.2.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.2.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.2.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.2.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.20.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.20.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.20.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1003,9 +990,8 @@
    "vision_model.encoder.layers.20.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.20.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.20.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.20.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.20.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.20.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.20.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.21.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.21.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.21.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1016,9 +1002,8 @@
    "vision_model.encoder.layers.21.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.21.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.21.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.21.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.21.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.21.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.21.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.22.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.22.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.22.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1029,9 +1014,8 @@
    "vision_model.encoder.layers.22.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.22.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.22.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.22.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.22.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.22.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.22.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.23.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.23.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.23.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1042,9 +1026,8 @@
    "vision_model.encoder.layers.23.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.23.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.23.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.23.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.23.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.23.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.23.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.24.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.24.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.24.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1055,9 +1038,8 @@
    "vision_model.encoder.layers.24.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.24.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.24.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.24.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.24.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.24.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.24.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.25.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.25.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.25.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1068,9 +1050,8 @@
    "vision_model.encoder.layers.25.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.25.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.25.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.25.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.25.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.25.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.25.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.26.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.26.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.26.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1081,9 +1062,8 @@
    "vision_model.encoder.layers.26.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.26.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.26.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.26.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.26.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.26.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.26.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.27.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.27.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.27.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1094,9 +1074,8 @@
    "vision_model.encoder.layers.27.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.27.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.27.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.27.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.27.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.27.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.27.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.28.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.28.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.28.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1107,9 +1086,8 @@
    "vision_model.encoder.layers.28.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.28.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.28.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.28.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.28.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.28.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.28.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.29.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.29.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.29.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1120,9 +1098,8 @@
    "vision_model.encoder.layers.29.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.29.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.29.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.29.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.29.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.29.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.29.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.3.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.3.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.3.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1133,9 +1110,8 @@
    "vision_model.encoder.layers.3.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.3.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.3.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.3.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.3.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.3.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.3.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.30.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.30.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.30.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1146,9 +1122,8 @@
    "vision_model.encoder.layers.30.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.30.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.30.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.30.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.30.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.30.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.30.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.31.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.31.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.31.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1159,9 +1134,8 @@
    "vision_model.encoder.layers.31.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.31.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.31.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.31.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.31.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.31.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.31.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.32.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.32.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.32.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1172,9 +1146,8 @@
    "vision_model.encoder.layers.32.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.32.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.32.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.32.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.32.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.32.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.32.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.33.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.33.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.33.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1185,9 +1158,8 @@
    "vision_model.encoder.layers.33.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.33.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.33.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.33.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.33.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.33.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.33.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.34.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.34.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.34.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1198,9 +1170,8 @@
    "vision_model.encoder.layers.34.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.34.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.34.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.34.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.34.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.34.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.34.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.35.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.35.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.35.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1211,9 +1182,8 @@
    "vision_model.encoder.layers.35.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.35.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.35.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.35.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.35.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.35.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.35.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.36.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.36.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.36.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1224,9 +1194,8 @@
    "vision_model.encoder.layers.36.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.36.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.36.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.36.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.36.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.36.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.36.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.37.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.37.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.37.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1237,9 +1206,8 @@
    "vision_model.encoder.layers.37.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.37.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.37.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.37.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.37.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.37.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.37.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.38.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.38.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.38.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1250,9 +1218,8 @@
    "vision_model.encoder.layers.38.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.38.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.38.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.38.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.38.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.38.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.38.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.4.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.4.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.4.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1263,9 +1230,8 @@
    "vision_model.encoder.layers.4.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.4.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.4.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.4.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.4.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.4.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.4.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.5.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.5.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.5.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1276,9 +1242,8 @@
    "vision_model.encoder.layers.5.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.5.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.5.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.5.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.5.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.5.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.5.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.6.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.6.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.6.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1289,9 +1254,8 @@
    "vision_model.encoder.layers.6.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.6.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.6.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.6.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.6.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.6.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.6.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.7.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.7.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.7.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1302,9 +1266,8 @@
    "vision_model.encoder.layers.7.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.7.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.7.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.7.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.7.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.7.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.7.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.8.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.8.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.8.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1315,9 +1278,8 @@
    "vision_model.encoder.layers.8.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.8.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.8.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.8.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.8.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.8.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.8.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.9.layer_norm1.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.9.layer_norm1.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.9.layer_norm2.bias": "pytorch_model-00001-of-00002.bin",
@ -1328,9 +1290,8 @@
    "vision_model.encoder.layers.9.mlp.fc2.weight": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.9.self_attn.projection.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.9.self_attn.projection.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.9.self_attn.q_bias": "pytorch_model-00001-of-00002.bin",
+    "vision_model.encoder.layers.9.self_attn.qkv.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.encoder.layers.9.self_attn.qkv.weight": "pytorch_model-00001-of-00002.bin",
-    "vision_model.encoder.layers.9.self_attn.v_bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.post_layernorm.bias": "pytorch_model-00001-of-00002.bin",
    "vision_model.post_layernorm.weight": "pytorch_model-00001-of-00002.bin"
  }
--- a/tokenizer_config.json
+++ b/tokenizer_config.json
@ -104,7 +104,6 @@
  "eos_token": "</s>",
  "extra_ids": 100,
  "model_max_length": 512,
-  "name_or_path": "google/flan-t5-xl",
  "pad_token": "<pad>",
  "processor_class": "Blip2Processor",
  "sp_model_kwargs": {},
Author	SHA1	Message	Date
JunnanLi	56fa169177	Update README.md	2023-03-18 02:23:31 +00:00
JunnanLi	7c74b4d09e	Update README.md	2023-03-18 02:16:03 +00:00
Niels Rogge	3d804dbd6d	Update README.md	2023-02-27 14:58:18 +00:00
Niels Rogge	096787f802	Update README.md	2023-02-10 08:36:54 +00:00
Niels Rogge	c147ed6239	Update README.md	2023-02-10 08:34:31 +00:00
Niels Rogge	5a253c24ff	Update README.md	2023-02-09 10:15:07 +00:00
Niels Rogge	8dd2e132f8	Create README.md	2023-02-09 10:14:56 +00:00
Niels Rogge	7855aa62e8	Upload Blip2ForConditionalGeneration	2023-02-08 20:44:35 +00:00
Niels Rogge	9fbd12cb97	Upload processor	2023-02-08 20:39:15 +00:00
Niels Rogge	55e353d249	Upload Blip2ForConditionalGeneration	2023-02-06 21:06:22 +00:00