Upload 4 files

app.py

@@ -0,0 +1,126 @@
+import torch
+import gradio as gr
+from transformers import AutoTokenizer
+from model import SmolLMForCausalLM, SmolLMConfig
+import os
+
+# 1. Configuration constants
+MODEL_CHECKPOINT = "model.pt" # Expects the model weights to be in this file
+TOKENIZER_ID = "HuggingFaceTB/SmolLM-135M" # Using the standard tokenizer
+DEVICE = "cpu" # HF Spaces free tier usually is CPU. Change to 'cuda' if GPU is available.
+
+# 2. Load Model and Tokenizer
+print("Loading tokenizer...")
+tokenizer = AutoTokenizer.from_pretrained(TOKENIZER_ID)
+
+print("Initializing model...")
+config = SmolLMConfig()
+model = SmolLMForCausalLM(config)
+
+# 3. Load Weights
+if os.path.exists(MODEL_CHECKPOINT):
+    print(f"Loading weights from {MODEL_CHECKPOINT}...")
+    try:
+        # Map location to CPU to be safe
+        checkpoint = torch.load(MODEL_CHECKPOINT, map_location=torch.device('cpu'))
+        
+        # Check if it's a full checkpoint (dict with 'model_state_dict') or just weights
+        if isinstance(checkpoint, dict) and 'model_state_dict' in checkpoint:
+            state_dict = checkpoint['model_state_dict']
+        else:
+            state_dict = checkpoint
+            
+        # Handle any prefix issues (e.g. if saved from compiled model with '_orig_mod.')
+        new_state_dict = {}
+        for k, v in state_dict.items():
+            if k.startswith("_orig_mod."):
+                new_state_dict[k[10:]] = v
+            else:
+                new_state_dict[k] = v
+                
+        model.load_state_dict(new_state_dict)
+        print("Weights loaded successfully.")
+    except Exception as e:
+        print(f"Error loading weights: {e}")
+        print("Running with initialized (random) weights for demonstration.")
+else:
+    print(f"Warning: {MODEL_CHECKPOINT} not found! Running with random weights.")
+
+model.to(DEVICE)
+model.eval()
+
+# 4. Generation Function
+def generate_text(prompt, max_new_tokens, temperature, top_k):
+    if not prompt:
+        return "Please enter a prompt."
+    
+    input_ids = tokenizer.encode(prompt, return_tensors="pt").to(DEVICE)
+    
+    # Text Generation Loop
+    # We implement a simple loop similar to the training script's generate function
+    # but added temperature and top-k sampling for better variety in the demo.
+    
+    curr_input_ids = input_ids
+    
+    with torch.no_grad():
+        for _ in range(int(max_new_tokens)):
+            # Get logits
+            logits = model(curr_input_ids)
+            next_token_logits = logits[:, -1, :]
+            
+            # Apply Temperature
+            if temperature > 0:
+                next_token_logits = next_token_logits / temperature
+            else:
+                # Greedy decoding if temperature is 0 (or very close)
+                # Just take argmax, but for code simplicity we'll let multinomial handle it with very high conf or Argmax
+                next_token_id = torch.argmax(next_token_logits, dim=-1).unsqueeze(0)
+                curr_input_ids = torch.cat([curr_input_ids, next_token_id], dim=1)
+                continue
+
+            # Apply Top-K
+            if top_k > 0:
+                v, _ = torch.topk(next_token_logits, min(top_k, next_token_logits.size(-1)))
+                next_token_logits[next_token_logits < v[:, [-1]]] = float('-inf')
+            
+            probs = torch.nn.functional.softmax(next_token_logits, dim=-1)
+            
+            # Sample
+            next_token_id = torch.multinomial(probs, num_samples=1)
+            curr_input_ids = torch.cat([curr_input_ids, next_token_id], dim=1)
+            
+            # optional: stop if EOS token is generated (if we had one defined and training used it)
+            # if next_token_id == tokenizer.eos_token_id:
+            #     break
+
+    output_text = tokenizer.decode(curr_input_ids[0].tolist(), skip_special_tokens=True)
+    return output_text
+
+# 5. Build Gradio Interface
+with gr.Blocks() as demo:
+    gr.Markdown("# SmolLM-135M Implementation Demo")
+    gr.Markdown("This is a demo of the 135M parameter transformer model trained from scratch.")
+    
+    with gr.Row():
+        with gr.Column():
+            prompt_input = gr.Textbox(label="Prompt", placeholder="Once upon a time...", lines=3)
+            with gr.Row():
+                max_tokens = gr.Slider(minimum=10, maximum=500, value=100, step=10, label="Max New Tokens")
+                temperature = gr.Slider(minimum=0.1, maximum=2.0, value=0.8, step=0.1, label="Temperature")
+            top_k = gr.Slider(minimum=1, maximum=100, value=40, step=1, label="Top-K")
+            generate_btn = gr.Button("Generate", variant="primary")
+            
+        with gr.Column():
+            output = gr.Textbox(label="Generated Text", lines=10)
+            
+    generate_btn.click(
+        fn=generate_text,
+        inputs=[prompt_input, max_tokens, temperature, top_k],
+        outputs=output
+    )
+    
+    gr.Markdown("### Note on inputs")
+    gr.Markdown("Because this model is small (135M) and trained on a specific dataset, it may not follow instructions like ChatGPT. It is best at completing text/stories.")
+
+if __name__ == "__main__":
+    demo.launch()

model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aba04700adc3a6ad2a4a89943dc7507c4393b5c9bab5eea07a6f8615278951e0
+size 538148429

model.py

@@ -0,0 +1,311 @@
+import math
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from dataclasses import dataclass
+from typing import Optional, Tuple
+
+@dataclass
+class SmolLMConfig:
+    """
+    Configuration class for SmolLM.
+    This holds all the hyperparameters that define the model architecture.
+    """
+    vocab_size: int = 49152  # Size of vocabulary (number of unique tokens)
+    hidden_size: int = 576   # Dimension of the embedding vectors
+    intermediate_size: int = 1536  # Dimension of the inner layer in the MLP
+    num_hidden_layers: int = 30    # Number of Transformer blocks (depth)
+    num_attention_heads: int = 9   # Number of heads for the query
+    num_key_value_heads: int = 3   # Number of heads for keys and values (GQA)
+    hidden_act: str = "silu"       # Activation function
+    max_position_embeddings: int = 2048 # Maximum sequence length
+    initializer_range: float = 0.02
+    rms_norm_eps: float = 1e-05
+    use_cache: bool = True
+    tie_word_embeddings: bool = True # Share weights between input embedding and output layer
+    rope_theta: float = 10000.0
+    
+    def __post_init__(self):
+        # Calculate dimension per head
+        self.head_dim = self.hidden_size // self.num_attention_heads
+        # Calculate how many Query heads share one Key/Value head (Grouped Query Attention)
+        self.num_key_value_groups = self.num_attention_heads // self.num_key_value_heads
+
+class RMSNorm(nn.Module):
+    """
+    Root Mean Square Layer Normalization (RMSNorm).
+    A simpler version of LayerNorm that re-scales inputs based on their RMS.
+    It stabilizes training and is used in Llama-based models instead of standard LayerNorm.
+    """
+    def __init__(self, dim: int, eps: float = 1e-5):
+        super().__init__()
+        self.eps = eps
+        self.weight = nn.Parameter(torch.ones(dim))
+
+    def _norm(self, x):
+        # Calculate RMS: sqrt(mean(x^2) + epsilon)
+        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)
+
+    def forward(self, x):
+        # Normalize and then scale by a learnable parameter
+        output = self._norm(x.float()).type_as(x)
+        return output * self.weight
+
+def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
+    """
+    Applies Rotary Positional Embeddings (RoPE) to queries and keys.
+    RoPE rotates the query and key vectors to inject relative positional information.
+    """
+    # q, k: [bs, num_heads, seq_len, head_dim]
+    # cos, sin: [seq_len, head_dim] or projected
+    
+    # Rotate function: [-x2, x1]
+    def rotate_half(x):
+        x1 = x[..., : x.shape[-1] // 2]
+        x2 = x[..., x.shape[-1] // 2 :]
+        return torch.cat((-x2, x1), dim=-1)
+
+    cos = cos.unsqueeze(0).unsqueeze(unsqueeze_dim) # [1, 1, seq_len, head_dim]
+    sin = sin.unsqueeze(0).unsqueeze(unsqueeze_dim)
+    
+    # Apply rotation: (x * cos) + (rotate_90(x) * sin)
+    q_embed = (q * cos) + (rotate_half(q) * sin)
+    k_embed = (k * cos) + (rotate_half(k) * sin)
+    return q_embed, k_embed
+
+class LlamaRotaryEmbedding(nn.Module):
+    """
+    Pre-computes the cosine and sine values for RoPE.
+    These are fixed values based on position indices, used to modulate Q and K.
+    """
+    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
+        super().__init__()
+        self.dim = dim
+        self.max_position_embeddings = max_position_embeddings
+        self.base = base
+        # Calculate inverse frequencies for the rotations
+        inv_freq = 1.0 / (self.base ** (torch.arange(0, self.dim, 2).float().to(device) / self.dim))
+        self.register_buffer("inv_freq", inv_freq, persistent=False)
+        self._set_cos_sin_cache(max_position_embeddings, device=device)
+
+    def _set_cos_sin_cache(self, seq_len, device):
+        self.max_seq_len_cached = seq_len
+        t = torch.arange(self.max_seq_len_cached, device=device, dtype=self.inv_freq.dtype)
+        freqs = torch.outer(t, self.inv_freq)
+        # Different from standard position embeddings, we concat freq to itself to cover both halves
+        emb = torch.cat((freqs, freqs), dim=-1)
+        self.register_buffer("cos_cached", emb.cos().to(dtype=torch.float32), persistent=False)
+        self.register_buffer("sin_cached", emb.sin().to(dtype=torch.float32), persistent=False)
+
+
+    def forward(self, x, seq_len):
+        if seq_len > self.max_seq_len_cached:
+            self._set_cos_sin_cache(seq_len=seq_len, device=x.device)
+        return (
+            self.cos_cached[:seq_len].to(dtype=x.dtype),
+            self.sin_cached[:seq_len].to(dtype=x.dtype),
+        )
+
+class LlamaMLP(nn.Module):
+    """
+    Feed-Forward Network (FFN) utilizing the SwiGLU activation.
+    Structure:
+    x -> GateProj -> SiLU \
+                           -> Multiply -> DownProj -> output
+    x -> UpProj_________/
+    """
+    def __init__(self, config: SmolLMConfig):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+        self.intermediate_size = config.intermediate_size
+        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
+        self.act_fn = nn.SiLU()
+
+    def forward(self, x):
+        # SwiGLU: (SiLU(Gate(x)) * Up(x)) -> Down(x)
+        down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
+        return down_proj
+
+class LlamaAttention(nn.Module):
+    """
+    Multi-Head Attention with Grouped Query Attention (GQA).
+    GQA uses fewer Key/Value heads than Query heads to save memory and KV cache during inference.
+    """
+    def __init__(self, config: SmolLMConfig):
+        super().__init__()
+        self.config = config
+        self.hidden_size = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = config.head_dim
+        self.num_key_value_heads = config.num_key_value_heads
+        self.num_key_value_groups = config.num_key_value_groups
+        self.max_position_embeddings = config.max_position_embeddings
+
+        self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
+        self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
+        self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
+        self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
+        
+        self.rotary_emb = LlamaRotaryEmbedding(self.head_dim, max_position_embeddings=self.max_position_embeddings)
+
+    def forward(self, x, position_ids=None, attention_mask=None):
+        bsz, q_len, _ = x.size()
+
+        # 1. Project inputs to Q, K, V
+        q = self.q_proj(x)
+        k = self.k_proj(x)
+        v = self.v_proj(x)
+
+        # 2. Reshape for multi-head attention
+        q = q.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        k = k.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+        v = v.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+
+        # 3. Apply Rotary Embeddings
+        cos, sin = self.rotary_emb(v, seq_len=q_len)
+        q, k = apply_rotary_pos_emb(q, k, cos, sin)
+
+        # 4. Handle GQA (Grouped Query Attention)
+        # If we have fewer KV heads than Q heads, we repeat K and V to match Q's dimensions
+        if self.num_key_value_groups > 1:
+            k = k[:, :, None, :, :].expand(bsz, self.num_key_value_heads, self.num_key_value_groups, q_len, self.head_dim).reshape(bsz, self.num_heads, q_len, self.head_dim)
+            v = v[:, :, None, :, :].expand(bsz, self.num_key_value_heads, self.num_key_value_groups, q_len, self.head_dim).reshape(bsz, self.num_heads, q_len, self.head_dim)
+
+        # 5. Scaled Dot Product Attention (Flash Attention / Memory Efficient Attention)
+        # We use PyTorch's optimized implementation which selects the best backend (FlashAttn, etc.)
+        # If passed an attention_mask, we might need to rely on the manual path if it's complex, 
+        # but for causal masking we can just use is_causal=True
+        
+        # NOTE: F.scaled_dot_product_attention expects 40D input: [batch, heads, seq, head_dim]
+        # Our q, k, v are already in that format after transpose.
+        
+        # If we have a mask that is NOT the causal mask (e.g. padding mask), we need to handle it.
+        # But for training from scratch with standard causal LM, we usually just need causal mask.
+        
+        dropout_p = 0.0 # Could add to config if desired
+        
+        # We need to broadcast the Mask if it is provided
+        if attention_mask is not None:
+             # Standard implementation if a custom mask is provided (rare for basic causal LM training)
+             attn_weights = torch.matmul(q, k.transpose(2, 3)) / math.sqrt(self.head_dim)
+             attn_weights = attn_weights + attention_mask
+             attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(q.dtype)
+             attn_output = torch.matmul(attn_weights, v)
+        else:
+             # Optimized path
+             attn_output = F.scaled_dot_product_attention(
+                q, k, v, 
+                attn_mask=None, 
+                dropout_p=dropout_p, 
+                is_causal=True
+            )
+        
+        # 6. Reshape back and apply output projection
+        attn_output = attn_output.transpose(1, 2).contiguous().reshape(bsz, q_len, self.hidden_size)
+        attn_output = self.o_proj(attn_output)
+        
+        return attn_output
+
+class LlamaDecoderLayer(nn.Module):
+    """
+    A single Transformer block.
+    Consists of:
+    1. Pre-Norm -> Attention -> Add Residual
+    2. Pre-Norm -> MLP (Feed Forward) -> Add Residual
+    """
+    def __init__(self, config: SmolLMConfig):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+        self.self_attn = LlamaAttention(config)
+        self.mlp = LlamaMLP(config)
+        self.input_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(self, x, position_ids=None, attention_mask=None):
+        residual = x
+        x = self.input_layernorm(x)
+        # Self Attention block
+        x = self.self_attn(x, position_ids=position_ids, attention_mask=attention_mask)
+        x = residual + x # Residual connection
+        
+        residual = x
+        x = self.post_attention_layernorm(x)
+        # MLP block
+        x = self.mlp(x)
+        x = residual + x # Residual connection
+        return x
+
+class SmolLMModel(nn.Module):
+    """
+    Main Transformer model (the "trunk").
+    Embeddings -> N x Decoder Layers -> Final Norm
+    """
+    def __init__(self, config: SmolLMConfig):
+        super().__init__()
+        self.config = config
+        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size)
+        self.layers = nn.ModuleList([LlamaDecoderLayer(config) for _ in range(config.num_hidden_layers)])
+        self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(self, input_ids):
+        # 1. Lookup Embeddings
+        x = self.embed_tokens(input_ids)
+        
+        seq_len = x.shape[1]
+        
+        # 2. Key Concept: Causal Mask
+        # We want the model to predict the NEXT token, so it shouldn't see future tokens.
+        # However, with F.scaled_dot_product_attention(is_causal=True), we don't need to pass an explicit mask
+        # unless dealing with padding.
+        # We pass None to allow the optimized attention to handle it.
+        mask = None
+        # mask = torch.full((seq_len, seq_len), float("-inf"), device=x.device)
+        # mask = torch.triu(mask, diagonal=1)
+        
+        # 3. Pass through all Transformer Layers
+        for layer in self.layers:
+            x = layer(x, attention_mask=mask)
+            
+        x = self.norm(x)
+        return x
+
+class SmolLMForCausalLM(nn.Module):
+    """
+    The full Causal Language Model.
+    Wraps the trunk (SmolLMModel) and adds the Language Model Head (Linear Layer) to project to accumulation logic.
+    """
+    def __init__(self, config: SmolLMConfig):
+        super().__init__()
+        self.model = SmolLMModel(config)
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+        
+        # Weight tying
+        if config.tie_word_embeddings:
+            self.lm_head.weight = self.model.embed_tokens.weight
+
+    def forward(self, input_ids):
+        x = self.model(input_ids)
+        logits = self.lm_head(x)
+        return logits
+
+def test_model():
+    config = SmolLMConfig()
+    print(f"Initializing SmolLM-135M with config: {config}")
+    
+    model = SmolLMForCausalLM(config)
+    print(f"Model keys: {model.state_dict().keys().__len__()}")
+    
+    # Test forward pass
+    dummy_input = torch.randint(0, config.vocab_size, (1, 32)) # Batch size 1, seq len 32
+    print(f"Running forward pass with input shape {dummy_input.shape}")
+    
+    logits = model(dummy_input)
+    print(f"Output shape: {logits.shape}") # Should be [1, 32, 49152]
+    
+    assert logits.shape == (1, 32, config.vocab_size)
+    print("Test passed!")
+
+if __name__ == "__main__":
+    test_model()

requirements.txt

@@ -0,0 +1,4 @@
+torch
+transformers
+gradio
+numpy