Upload DeepSeekV2Lite DenseMixer model

DeepSeekV2Lite model with DenseMixer architecture

download_original.py

@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+"""
+Download original DeepSeek-V2-Lite model and copy to our directory
+"""
+
+import os
+from huggingface_hub import snapshot_download
+import shutil
+
+# Download original model
+print("Downloading original DeepSeek-V2-Lite model...")
+original_path = snapshot_download(
+    repo_id="deepseek-ai/DeepSeek-V2-Lite",
+    cache_dir="/tmp/deepseek_download"
+)
+
+print(f"Downloaded to: {original_path}")
+
+# Target directory
+target_dir = "/mnt/weka/home/shibo.hao/feng/code/junxia/deepseekv2lite"
+
+# Files to copy (only model weights, keep our custom files)
+files_to_copy = [
+    "model-00001-of-000004.safetensors",
+    "model-00002-of-000004.safetensors", 
+    "model-00003-of-000004.safetensors",
+    "model-00004-of-000004.safetensors",
+    "model.safetensors.index.json"
+]
+
+# Copy files
+print("Copying model weight files...")
+for filename in files_to_copy:
+    src = os.path.join(original_path, filename)
+    dst = os.path.join(target_dir, filename)
+    
+    if os.path.exists(src):
+        print(f"Copying {filename}...")
+        shutil.copy2(src, dst)
+        # Check file size
+        size = os.path.getsize(dst)
+        print(f"  Size: {size / (1024**3):.2f} GB")
+    else:
+        print(f"Warning: {filename} not found in original model")
+
+print("Done! Model weights copied successfully.")

model-00001-of-000004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:258811fe017b955eb542325c73f8cd8fc2f77f7c5c27dc1721877212182bffe4
-size 135
+oid sha256:0d7e9f39bde40111a4c0f390b87497dce4565cf578d916395e6b2c7851f1e8da
+size 8594887408

model-00002-of-000004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:46264772a9631656d6f8ff201943336f45d298e2dade91cbf3bf87f61dbc10eb
-size 135
+oid sha256:e0656832b0d594b4a64cad40ff8465231de6ed44c854f74f6b217797660aa4bb
+size 8591757448

model-00003-of-000004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e5c3a29ec4c6a815eaf1fbff20ca641ce9f3dc030fd7f7d5ea6d4962605322f7
-size 135
+oid sha256:843ec689624f3a520526e040f0326c4dc9865e8172942ca98a084fe136fdb21a
+size 8590718520

model-00004-of-000004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:900e5ae475e2105613495a483a4aca219d3de2c07757e01d6db3ea4abe682790
-size 135
+oid sha256:cfb51658f67cedfbbc4d62ad14187830ceec8ee82c788c5f718feea98905ef31
+size 5636263200