From 8e6e30ef89a46b7cc780c2c5398eef95a8780096 Mon Sep 17 00:00:00 2001
From: yuluoyun <1731396519@qq.com>
Date: Wed, 10 Jan 2024 12:14:44 +0800
Subject: [PATCH] add monkey.py

---
 vlmeval/config.py       |  1 +
 vlmeval/vlm/__init__.py |  1 +
 vlmeval/vlm/monkey.py   | 43 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 45 insertions(+)
 create mode 100644 vlmeval/vlm/monkey.py

diff --git a/vlmeval/config.py b/vlmeval/config.py
index e55ed056..b1430e5a 100644
--- a/vlmeval/config.py
+++ b/vlmeval/config.py
@@ -43,6 +43,7 @@
     'cogvlm-grounding-generalist':partial(CogVlm, name='cogvlm-grounding-generalist',tokenizer_name ='lmsys/vicuna-7b-v1.5'),
     'cogvlm-chat':partial(CogVlm, name='cogvlm-chat',tokenizer_name ='lmsys/vicuna-7b-v1.5'),
     'sharedcaptioner':partial(SharedCaptioner, model_path='Lin-Chen/ShareCaptioner'),
+    'monkey':partial(Monkey, model_path='echo840/Monkey'),
 }
 
 api_models = {
diff --git a/vlmeval/vlm/__init__.py b/vlmeval/vlm/__init__.py
index e5172a83..117743e8 100644
--- a/vlmeval/vlm/__init__.py
+++ b/vlmeval/vlm/__init__.py
@@ -15,3 +15,4 @@
 from .llava_xtuner import LLaVA_XTuner
 from .cogvlm import CogVlm
 from .sharedcaptioner import SharedCaptioner
+from .monkey import Monkey
diff --git a/vlmeval/vlm/monkey.py b/vlmeval/vlm/monkey.py
new file mode 100644
index 00000000..6dc6c71d
--- /dev/null
+++ b/vlmeval/vlm/monkey.py
@@ -0,0 +1,43 @@
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import warnings
+import os.path as osp
+from vlmeval.smp import isimg
+import re
+
+class Monkey:
+
+    INSTALL_REQ = False
+
+    def __init__(self, model_path='echo840/Monkey', **kwargs):
+        assert model_path is not None
+        self.model_path = model_path
+        self.tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+        self.model = AutoModelForCausalLM.from_pretrained(model_path, device_map='cuda', trust_remote_code=True).eval()
+        self.kwargs = kwargs
+        warnings.warn(f"Following kwargs received: {self.kwargs}, will use as generation config. ")
+        torch.cuda.empty_cache()
+    
+    def generate(self, image_path, prompt, dataset=None):
+        cur_prompt = f'<img>{image_path}</img> {prompt} Answer: '
+        input_ids = self.tokenizer(cur_prompt, return_tensors='pt', padding='longest')
+        attention_mask = input_ids.attention_mask
+        input_ids = input_ids.input_ids
+        
+        output_ids = self.model.generate(
+                input_ids=input_ids.cuda(),
+                attention_mask=attention_mask.cuda(),
+                do_sample=False,
+                num_beams=1,
+                max_new_tokens=512,
+                min_new_tokens=1,
+                length_penalty=3,
+                num_return_sequences=1,
+                output_hidden_states=True,
+                use_cache=True,
+                pad_token_id=self.tokenizer.eod_id,
+                eos_token_id=self.tokenizer.eod_id,
+            )
+        response = self.tokenizer.decode(output_ids[0][input_ids.size(1):].cpu(), skip_special_tokens=True).strip()
+        
+        return response