Upload 3 files

.gitattributes

@@ -38,3 +38,5 @@ gen_img_embeds.rknn filter=lfs diff=lfs merge=lfs -text
 image_decode.rknn filter=lfs diff=lfs merge=lfs -text
 lm_head.rknn filter=lfs diff=lfs merge=lfs -text
 vision_encoder.rknn filter=lfs diff=lfs merge=lfs -text
+generated.png filter=lfs diff=lfs merge=lfs -text
+language_model.rkllm filter=lfs diff=lfs merge=lfs -text

language_model.rkllm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:936e2c126b39ea35e60741e09db18c2706b07a7169da1f51a865686f0e0720f3
+size 3330314348

run_rkllm.py

@@ -5,7 +5,7 @@ import os
 os.environ["RKLLM_LOG_LEVEL"] = "1"
 import numpy as np
 import onnxruntime as real_ort
-import ztu_somemodelruntime_rknn2 as ort
+import ztu_somemodelruntime_rknnlite2 as ort
 from tokenizers import Tokenizer
 import cv2
 import tqdm
@@ -19,13 +19,13 @@ onnx_model_path = f"{model_path}"
 tokenizer = Tokenizer.from_file(f"{model_path}/tokenizer.json")
 # np.random.seed(0)
 
-# image = None
-# prompt = "A stunning princess from kabul in red, white traditional clothing, blue eyes, brown hair"
-# mode = "t2i" # 文本生成图片, 反过来是it2t -> 图片/文本生成文本
+image = None
+prompt = "A stunning princess from kabul in red, white traditional clothing, blue eyes, brown hair"
+mode = "t2i" # 文本生成图片, 反过来是it2t -> 图片/文本生成文本
 
-image = "./test.jpg"
-prompt = "仔细描述这张图片。"
-mode = "it2t"
+# image = "./test.jpg"
+# prompt = "仔细描述这张图片。"
+# mode = "it2t"
 
 tempature = 0.7