add cpuinfo

app.py

@@ -5,6 +5,8 @@ import time
 from threading import Thread
 import sys
 import os
+os.environ["BNB_CUDA_VERSION"] = "0" # Forces bitsandbytes to recognize no GPU
+os.environ["OMP_NUM_THREADS"] = "1" # Prevents race conditions in custom CPU kernels
 
 # Model configuration
 if len(sys.argv) > 1 and os.path.exists(sys.argv[1]):
@@ -17,24 +19,48 @@ else:
 tokenizer = None
 model = None
 
+import platform
+import subprocess
+import cpuinfo  # Optional: 'pip install py-cpuinfo' is better if you can add it
+
 def load_model():
-    """Load the Smol LLM model and tokenizer"""
+    """Load the Smol LLM model and tokenizer with hardware detection"""
     global tokenizer, model
     try:
+        print("--- Hardware Audit ---")
+        print(f"Processor: {platform.processor()}")
+        print(f"Machine: {platform.machine()}")
+        
+        # Check for CPU Flags (Instruction Sets)
+        try:
+            # For Linux-based Cloud environments
+            cpu_flags = subprocess.check_output("lscpu", shell=True).decode()
+            print("Instruction sets found:")
+            for flag in ["avx512", "avx2", "avx", "fma", "amx"]:
+                if flag in cpu_flags.lower():
+                    print(f"  ✅ {flag.upper()} supported")
+                else:
+                    print(f"  ❌ {flag.upper()} NOT found")
+        except Exception as e:
+            print(f"Could not check CPU flags: {e}")
+        
+        print(f"PyTorch version: {torch.__version__}")
         print(f"Loading model: {MODEL_NAME}")
+        print("----------------------")
+
         tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, use_fast=True)
-        tokenizer.padding_side = "left" # Ensure consistency
+        tokenizer.padding_side = "left"
         if tokenizer.pad_token is None:
             tokenizer.pad_token = tokenizer.eos_token
+
         model = AutoModelForCausalLM.from_pretrained(
             MODEL_NAME,
+            # If the cloud lacks AVX512, float32 is safer but slower
             dtype=torch.float32,
-            device_map={"": "cpu"}
+            device_map={"": "cpu"},
+            low_cpu_mem_usage=True
         )
-        
-        if tokenizer.pad_token is None:
-            tokenizer.pad_token = tokenizer.eos_token
-            
+
         return "✅ Model loaded successfully!"
     except Exception as e:
         return f"❌ Error loading model: {str(e)}"

requirements.txt

@@ -1,17 +1,28 @@
 accelerate==1.12.0
 aiofiles==23.2.1
+aiohappyeyeballs==2.6.1
+aiohttp==3.13.2
+aiosignal==1.4.0
 annotated-doc==0.0.4
 annotated-types==0.7.0
 anyio==4.12.0
+async-timeout==5.0.1
+attrs==25.4.0
+Authlib==1.6.6
 bitsandbytes==0.49.0
 brotli==1.2.0
 certifi==2025.11.12
+cffi==2.0.0
 charset-normalizer==3.4.4
 click==8.3.1
+cryptography==46.0.3
+datasets==4.4.2
+dill==0.4.0
 exceptiongroup==1.3.1
 fastapi==0.125.0
 ffmpy==1.0.0
 filelock==3.20.1
+frozenlist==1.8.0
 fsspec==2025.10.0
 gradio==5.9.0
 gradio_client==1.5.2
@@ -27,6 +38,8 @@ markdown-it-py==4.0.0
 MarkupSafe==2.1.5
 mdurl==0.1.2
 mpmath==1.3.0
+multidict==6.7.0
+multiprocess==0.70.18
 networkx==3.4.2
 numpy==2.2.6
 nvidia-cublas-cu12==12.8.4.1
@@ -48,7 +61,11 @@ orjson==3.11.5
 packaging==25.0
 pandas==2.3.3
 pillow==11.3.0
+propcache==0.4.1
 psutil==5.9.8
+py-cpuinfo==9.0.0
+pyarrow==22.0.0
+pycparser==2.23
 pydantic==2.12.4
 pydantic_core==2.41.5
 pydub==0.25.1
@@ -66,13 +83,14 @@ safetensors==0.7.0
 semantic-version==2.10.0
 shellingham==1.5.4
 six==1.17.0
+spaces==0.45.0
 starlette==0.50.0
 sympy==1.14.0
 tokenizers==0.22.1
 tomlkit==0.13.3
 torch==2.9.1
 tqdm==4.67.1
-transformers @ git+https://github.com/708-145/transformers.git@hirma
+transformers @ git+https://github.com/708-145/transformers.git@4aacbc4f98e1c802cfeb7602a95c594321f1d360
 triton==3.5.1
 typer==0.20.0
 typer-slim==0.20.0
@@ -82,3 +100,5 @@ tzdata==2025.3
 urllib3==2.6.2
 uvicorn==0.38.0
 websockets==14.2
+xxhash==3.6.0
+yarl==1.22.0