BoyaWu10 commited on
Commit
ea0b235
·
verified ·
1 Parent(s): 4fb70c3

Upload model weights

Browse files
Files changed (4) hide show
  1. README.md +14 -0
  2. config.json +38 -0
  3. mm_projector.bin +3 -0
  4. trainer_state.json +0 -0
README.md CHANGED
@@ -1,3 +1,17 @@
1
  ---
 
2
  license: apache-2.0
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ inference: false
3
  license: apache-2.0
4
  ---
5
+
6
+ # Model Card
7
+
8
+ Bunny is a family of lightweight multimodal models.
9
+
10
+ Bunny-pretrain-qwen1.5-1.8b-siglip is the pretrained weights for [bunny-qwen1.5-1.8b-siglip-lora](https://huggingface.co/BoyaWu10/bunny-qwen1.5-1.8b-siglip-lora), which leverages Qwen1.5-1.8B as the language model backbone and SigLIP as the vision encoder.
11
+ It is pretrained on LAION-2M.
12
+
13
+ More details about this model can be found in [GitHub](https://github.com/BAAI-DCAI/Bunny).
14
+
15
+ # License
16
+ This project utilizes certain datasets and checkpoints that are subject to their respective original licenses. Users must comply with all terms and conditions of these original licenses.
17
+ The content of this project itself is licensed under the Apache license 2.0.
config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "bunny-pretrain-qwen1.5-1.8b-siglip",
3
+ "architectures": [
4
+ "Qwen2ForCausalLM"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 151643,
8
+ "eos_token_id": 151643,
9
+ "freeze_mm_mlp_adapter": false,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 2048,
12
+ "image_aspect_ratio": "square",
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 5504,
15
+ "max_position_embeddings": 32768,
16
+ "max_window_layers": 21,
17
+ "mm_hidden_size": 1152,
18
+ "mm_projector_lr": null,
19
+ "mm_projector_type": "mlp2x_gelu",
20
+ "mm_vision_tower": "google/siglip-so400m-patch14-384",
21
+ "model_type": "bunny-qwen2",
22
+ "num_attention_heads": 16,
23
+ "num_hidden_layers": 24,
24
+ "num_key_value_heads": 16,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_theta": 1000000.0,
27
+ "sliding_window": 32768,
28
+ "tie_word_embeddings": false,
29
+ "tokenizer_model_max_length": 2048,
30
+ "tokenizer_padding_side": "right",
31
+ "torch_dtype": "bfloat16",
32
+ "transformers_version": "4.38.2",
33
+ "tune_mm_mlp_adapter": true,
34
+ "use_cache": true,
35
+ "use_mm_proj": true,
36
+ "use_sliding_window": false,
37
+ "vocab_size": 151936
38
+ }
mm_projector.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eeac5522482d994d673c482e0f5df22ce0154a28f3aab4519505f06a04410f67
3
+ size 13117496
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff