moved both cbow and qwen to cache HF dataset

.dockerignore

@@ -31,4 +31,6 @@ models/
 
 # Docker or Space-specific
 docker-compose.yaml
-Caddyfile
+Caddyfile
+qwen_model/
+models/

.gitignore

@@ -2,3 +2,4 @@ __pycache__/
 *.png
 **/artifacts/
 models/
+qwen_model/

Dockerfile

@@ -22,13 +22,21 @@ RUN mkdir -p /models/clip && \
 RUN python3 -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('bert-base-uncased').save_pretrained('/models/bert-tokenizer')"
 RUN python3 -c "from transformers import CLIPProcessor; CLIPProcessor.from_pretrained('openai/clip-vit-base-patch32').save_pretrained('/models/clip')"
 
+# RUN mkdir -p /models/cbow && \
+#     python3 -c "import gensim.downloader as api; model = api.load('glove-twitter-200'); model.save('/models/cbow_model.kv')"
 RUN mkdir -p /models/cbow && \
-    python3 -c "import gensim.downloader as api; model = api.load('glove-twitter-200'); model.save('/models/cbow_model.kv')"
+    wget https://huggingface.co/datasets/ClemSummer/cbow-model-cache/resolve/main/cbow_model.kv -O /models/cbow/cbow_model.kv && \
+    wget https://huggingface.co/datasets/ClemSummer/cbow-model-cache/resolve/main/cbow_model.kv.vectors.npy -O /models/cbow/cbow_model.kv.vectors.npy
 
+# RUN mkdir -p /models/qwen && \
+#     python3 -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('Qwen/Qwen3-0.6B-Base').save_pretrained('/models/qwen')"
+# RUN python3 -c "from transformers import AutoModelForCausalLM; AutoModelForCausalLM.from_pretrained('Qwen/Qwen3-0.6B-Base').save_pretrained('/models/qwen')"
 RUN mkdir -p /models/qwen && \
-    python3 -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('Qwen/Qwen3-0.6B-Base').save_pretrained('/models/qwen')"
-RUN python3 -c "from transformers import AutoModelForCausalLM; AutoModelForCausalLM.from_pretrained('Qwen/Qwen3-0.6B-Base').save_pretrained('/models/qwen')"
+    python3 -c "from transformers import AutoTokenizer; \
+                AutoTokenizer.from_pretrained('ClemSummer/qwen-model-cache', trust_remote_code=True).save_pretrained('/models/qwen')"
 
+RUN python3 -c "from transformers import AutoModelForCausalLM; \
+                AutoModelForCausalLM.from_pretrained('ClemSummer/qwen-model-cache', trust_remote_code=True).save_pretrained('/models/qwen')"
 EXPOSE 7860
 
 # Install curl if it's not already installed

cbow_logic.py

@@ -7,7 +7,7 @@ import shlex
 
 
 class MeaningCalculator:
-    def __init__(self, model_path: str = "/models/cbow_model.kv"):
+    def __init__(self, model_path: str = "/models/cbow/cbow_model.kv"):
         if not os.path.exists(model_path):
             raise FileNotFoundError(f"Model not found at: {model_path}")
         self.model = gensim.models.KeyedVectors.load(model_path, mmap='r')