import torch import torch.nn.functional as F import torch.distributions as dists import transformers from transformers import AutoTokenizer, AutoModelForCausalLM from peft import PeftModel, PeftConfig import numpy as np import random import time import os from typing import List, Dict, Optional, Tuple, Iterator, Set import gradio as gr import gc # Suppress some Hugging Face warnings os.environ["TOKENIZERS_PARALLELISM"] = "false" # Import necessary model classes # Assuming these custom classes are in the correct path from model_cache.llada.modeling_llada import LLaDAModelLM from model_cache.llada.configuration_llada import LLaDAConfig # --- Helper Functions (Unchanged) --- def set_seed(seed): torch.manual_seed(seed); random.seed(seed); np.random.seed(seed); if torch.cuda.is_available(): torch.cuda.manual_seed_all(seed); torch.backends.cudnn.deterministic = True; torch.backends.cudnn.benchmark = False def create_full_block_attention_mask(prompt_length, max_length, block_size, device=None, dtype=None): if dtype is None: dtype = torch.bfloat16 attention_mask = torch.full((1, 1, max_length, max_length), -torch.inf, device=device, dtype=dtype) attention_mask[:, :, :prompt_length, :prompt_length] = 0 remaining_length = max_length - prompt_length num_blocks = (remaining_length + block_size - 1) // block_size for b in range(num_blocks): block_start = prompt_length + b * block_size; block_end = min(prompt_length + (b + 1) * block_size, max_length) attention_mask[:, :, block_start:block_end, :prompt_length] = 0 for prev_b in range(b): prev_start = prompt_length + prev_b * block_size; prev_end = min(prompt_length + (prev_b + 1) * block_size, max_length) attention_mask[:, :, block_start:block_end, prev_start:prev_end] = 0 attention_mask[:, :, block_start:block_end, block_start:block_end] = 0 return attention_mask def extract_attention_mask(full_mask, start_pos, input_length, cache_length): end_pos = start_pos + input_length; total_length = cache_length + input_length extracted_mask = torch.full((1, 1, input_length, total_length), -torch.inf, device=full_mask.device, dtype=full_mask.dtype) extracted_mask[:, :, :, :cache_length] = full_mask[:, :, start_pos:end_pos, :cache_length] extracted_mask[:, :, :, cache_length:] = full_mask[:, :, start_pos:end_pos, start_pos:end_pos] return extracted_mask def top_p_logits(logits, top_p=None): sorted_logits, sorted_indices = torch.sort(logits, descending=True) cumulative_probs = torch.cumsum(F.softmax(sorted_logits, dim=-1), dim=-1) sorted_indices_to_remove = cumulative_probs > top_p sorted_indices_to_remove[..., 1:] = sorted_indices_to_remove[..., :-1].clone() sorted_indices_to_remove[..., 0] = 0 mask = torch.zeros_like(logits, dtype=torch.bool, device=logits.device) mask = mask.scatter_(-1, sorted_indices, sorted_indices_to_remove) logits = logits.masked_fill(mask, torch.finfo(logits.dtype).min) return logits def top_k_logits(logits, top_k=None): top_k = min(top_k, logits.size(-1)) indices_to_remove = logits < torch.topk(logits, top_k)[0][..., -1, None] logits = logits.masked_fill(indices_to_remove, torch.finfo(logits.dtype).min) return logits def sample_tokens(logits, temperature=0.0, top_p=None, top_k=None, margin_confidence=False, neg_entropy=False): if temperature > 0: logits = logits / temperature if top_p is not None and top_p < 1: logits = top_p_logits(logits, top_p) if top_k is not None: logits = top_k_logits(logits, top_k) probs = torch.softmax(logits, dim=-1) if temperature > 0: try: x0 = dists.Categorical(probs=probs).sample() initial_confidence = torch.gather(probs, -1, x0.unsqueeze(-1)).squeeze(-1) except: initial_confidence, x0 = probs.max(dim=-1) else: initial_confidence, x0 = probs.max(dim=-1) confidence = initial_confidence.clone() if margin_confidence: sorted_probs, _ = torch.sort(probs, dim=-1, descending=True) confidence = sorted_probs[:, 0] - sorted_probs[:, 1] if neg_entropy: epsilon = 1e-10 confidence = torch.sum(probs * torch.log(probs + epsilon), dim=-1) return confidence, x0, initial_confidence class D2FInference: CSS = """ .gradio-container { font-family: -apple-system, BlinkMacSystemFont, sans-serif; } .model-header { font-size: 1.2em; font-weight: bold; margin-bottom: 10px; padding: 8px; border-radius: 5px; text-align: center; } .d2f-header { background-color: #DBEAFE; color: #1E40AF; } .llama-header { background-color: #FEF3C7; color: #92400E; } .stats-container { padding: 15px; border: 1px solid #10B981; border-radius: 8px; background-color: #F0FDF4; margin-top: 10px; margin-bottom: 20px; } .output-textbox textarea { font-size: 1.5em !important; line-height: 1.6 !important; height: 70vh !important; overflow-y: auto !important; } """ def __init__(self, **kwargs): print("Initializing D2F-LLaDA model...") self.device = torch.device(kwargs.get("device", "cuda:3") if torch.cuda.is_available() else "cpu") self.__dict__.update(kwargs) if self.dtype == "bfloat16" and torch.cuda.is_bf16_supported(): self.target_dtype = torch.bfloat16 elif self.dtype == "float16": self.target_dtype = torch.float16 else: self.target_dtype = torch.float32 self._setup_model(self.pretrained_path, self.lora_path) print("D2F-LLaDA model and tokenizer setup complete.") def _setup_model(self, pretrained_path, lora_path): config = LLaDAConfig.from_pretrained(pretrained_path) self.model = LLaDAModelLM.from_pretrained(pretrained_path, config=config, torch_dtype=self.target_dtype).eval() self.model = PeftModel.from_pretrained(self.model, lora_path) self.model = self.model.to(self.device) self.tokenizer = AutoTokenizer.from_pretrained(pretrained_path) if self.tokenizer.pad_token is None: self.tokenizer.pad_token = self.tokenizer.eos_token def _apply_chat_template(self, prompt): chat_history = [{"role": "user", "content": prompt}] return self.tokenizer.apply_chat_template(chat_history, tokenize=False, add_generation_prompt=True) def _update_block_completion_states(self, block_states, decoded_token_threshold): for block_id in sorted(block_states.keys()): decoded_tokens = block_states[block_id]['total_masks'] - block_states[block_id]['mask_count'] if block_states[block_id]['total_masks'] > 0: decode_ratio = decoded_tokens / block_states[block_id]['total_masks'] if decode_ratio >= decoded_token_threshold: if (next_block_id := block_id + 1) in block_states: block_states[next_block_id]['is_complete'] = True @torch.inference_mode() def stream( self, prompt_text: str, max_new_tokens: int, block_size: int, block_add_threshold: float, decoded_token_threshold: float, skip_threshold: float ) -> Iterator[Tuple[str, str]]: start_time = time.time() input_ids = self.tokenizer(self._apply_chat_template(prompt_text), return_tensors="pt").input_ids.to(self.device) prompt_length = input_ids.shape[1] full_attention_mask = create_full_block_attention_mask(prompt_length, self.max_length, block_size, self.device, self.target_dtype) x_t = input_ids block_states = {0: {'start_pos': 0, 'end_pos': prompt_length, 'mask_count': 0, 'total_masks': prompt_length, 'state': 'to_cache', 'is_complete': True}} past_key_values, current_blocks, step, eos_detected, cache_length = None, 0, 0, False, 0 yield "", None tokens_generated = 0 while True: step += 1 updated_block_ids = set() if len(block_states) - 1 < (max_new_tokens // block_size) and not eos_detected: last_block_id = max(block_states.keys()) progress_ratio = (block_states[last_block_id]['total_masks'] - block_states[last_block_id]['mask_count']) / block_states[last_block_id]['total_masks'] if block_states[last_block_id]['total_masks'] > 0 else 1.0 if progress_ratio >= block_add_threshold: new_block_id = last_block_id + 1; new_start_pos = x_t.shape[1] if new_start_pos + block_size <= self.max_length: x_t = torch.cat([x_t, torch.full((1, block_size), self.mask_token_id, device=self.device, dtype=torch.long)], dim=1) block_states[new_block_id] = {'start_pos': new_start_pos, 'end_pos': new_start_pos + block_size, 'mask_count': block_size, 'total_masks': block_size, 'state': 'active', 'is_complete': False} current_blocks += 1 self._update_block_completion_states(block_states, decoded_token_threshold) if (x_t == self.mask_token_id).sum() == 0 and current_blocks == 0: break blocks_to_cache = [bid for bid, state in block_states.items() if state['state'] == 'to_cache'] update_kvcache = 0 if blocks_to_cache: start_pos, end_pos = block_states[min(blocks_to_cache)]['start_pos'], block_states[max(blocks_to_cache)]['end_pos'] update_kvcache = end_pos - start_pos; input_seq, process_start_pos = x_t[:, start_pos:], start_pos else: active_blocks = [bid for bid, state in block_states.items() if state['state'] == 'active' and state['start_pos'] >= cache_length] if not active_blocks: break start_pos = min(block_states[bid]['start_pos'] for bid in active_blocks); input_seq, process_start_pos = x_t[:, start_pos:], start_pos if input_seq.shape[1] == 0: break attention_mask = extract_attention_mask(full_mask=full_attention_mask, start_pos=process_start_pos, input_length=input_seq.shape[1], cache_length=cache_length) outputs = self.model(input_seq, attention_bias=attention_mask, past_key_values=past_key_values, use_cache=True, update_kvcache=update_kvcache + cache_length) if update_kvcache > 0: past_key_values = outputs.past_key_values for bid in blocks_to_cache: block_states[bid]['state'] = 'in_cache' blocks_to_deactivate = [] for block_id, state in block_states.items(): if state['state'] != 'active': continue block_mask_locs = (x_t[0, state['start_pos']:state['end_pos']] == self.mask_token_id).nonzero().squeeze(-1) if block_mask_locs.numel() == 0: blocks_to_deactivate.append(block_id) continue logit_offset = state['start_pos'] - process_start_pos block_mask_logits = outputs.logits[:, logit_offset + block_mask_locs, :] _, x0, initial_confidence = sample_tokens(block_mask_logits.squeeze(0), self.temperature, self.top_p, self.top_k) all_indices = (initial_confidence > skip_threshold).nonzero().squeeze(-1) if state['is_complete'] and all_indices.numel() == 0 and block_mask_logits.numel() > 0: all_indices = torch.tensor([torch.argmax(initial_confidence)], device=self.device) if all_indices.numel() > 0: updated_block_ids.add(block_id) positions_to_update = state['start_pos'] + block_mask_locs[all_indices] x_t[0, positions_to_update] = x0[all_indices] state['mask_count'] -= all_indices.numel() tokens_generated += all_indices.numel() if self.tokenizer.eos_token_id in x0[all_indices]: eos_detected = True if state['mask_count'] == 0: blocks_to_deactivate.append(block_id) for bid in blocks_to_deactivate: if block_states[bid]['state'] == 'active' and all(block_states.get(i, {}).get('state') != 'active' for i in range(bid)): block_states[bid]['state'] = 'to_cache' current_blocks -= 1 if update_kvcache > 0: cache_length += update_kvcache generated_ids = x_t[0, prompt_length:] valid_ids = generated_ids[generated_ids != self.mask_token_id] live_text = self.tokenizer.decode(valid_ids, skip_special_tokens=True) yield live_text, None total_time = time.time() - start_time final_generated_ids = x_t[0, prompt_length:] eos_positions = (final_generated_ids == self.tokenizer.eos_token_id).nonzero() if eos_positions.numel() > 0: final_generated_ids = final_generated_ids[:eos_positions[0, 0] + 1] final_text = self.tokenizer.decode(final_generated_ids, skip_special_tokens=True) tokens_incl_eos = len(final_generated_ids) tokens_per_second = tokens_incl_eos / total_time if total_time > 0 else 0 stats = { "total_time": total_time, "tokens_generated": tokens_incl_eos, "tokens_per_second": tokens_per_second } if past_key_values is not None: del past_key_values del full_attention_mask torch.cuda.empty_cache() yield final_text, stats class LlamaInference: def __init__(self, **kwargs): print("Initializing LLaMA model...") self.device = torch.device(kwargs.get("device", "cuda:4") if torch.cuda.is_available() else "cpu") self.__dict__.update(kwargs) self._setup_model(self.model_id) print("LLaMA model and tokenizer setup complete.") def _setup_model(self, model_id): print(f"Loading LLaMA model {model_id} on {self.device}...") self.tokenizer = AutoTokenizer.from_pretrained(model_id) self.model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16 if torch.cuda.is_bf16_supported() else torch.float16, device_map=self.device ).eval() if self.tokenizer.eos_token is None: self.tokenizer.eos_token = "" if self.tokenizer.pad_token is None: self.tokenizer.pad_token = self.tokenizer.eos_token def _apply_chat_template(self, prompt): chat_history = [{"role": "user", "content": prompt}] return self.tokenizer.apply_chat_template(chat_history, tokenize=False, add_generation_prompt=True) @torch.inference_mode() def stream( self, prompt_text: str, max_new_tokens: int, temperature: float = 0.0, top_p: float = 0.9, top_k: int = None ) -> Iterator[Tuple[str, str]]: start_time = time.time() formatted_prompt = self._apply_chat_template(prompt_text) input_ids = self.tokenizer(formatted_prompt, return_tensors="pt").input_ids.to(self.device) prompt_length = input_ids.shape[1] yield "", None tokens_generated = 0 current_input_ids = input_ids.clone() for i in range(max_new_tokens): with torch.no_grad(): outputs = self.model(current_input_ids, use_cache=True) next_token_logits = outputs.logits[:, -1, :] if temperature > 0: next_token_logits = next_token_logits / temperature if top_p is not None and top_p < 1: next_token_logits = top_p_logits(next_token_logits, top_p) if top_k is not None: next_token_logits = top_k_logits(next_token_logits, top_k) probs = torch.softmax(next_token_logits, dim=-1) next_token = torch.multinomial(probs, num_samples=1) else: next_token = torch.argmax(next_token_logits, dim=-1, keepdim=True) current_input_ids = torch.cat([current_input_ids, next_token], dim=-1) tokens_generated += 1 if next_token[0, 0].item() == self.tokenizer.eos_token_id: break generated_text = self.tokenizer.decode( current_input_ids[0, prompt_length:], skip_special_tokens=True ) yield generated_text, None del outputs total_time = time.time() - start_time tokens_per_second = tokens_generated / total_time if total_time > 0 else 0 final_text = self.tokenizer.decode(current_input_ids[0, prompt_length:], skip_special_tokens=True) stats = { "total_time": total_time, "tokens_generated": tokens_generated, "tokens_per_second": tokens_per_second } del current_input_ids torch.cuda.empty_cache() yield final_text, stats # --- Comparison Helper Functions --- def create_comparison_html(d2f_results, llama_results): d_tokens = d2f_results["tokens_generated"] d_time = d2f_results["total_time"] d_tokens_per_sec = d2f_results["tokens_per_second"] a_tokens = llama_results["tokens_generated"] a_time = llama_results["total_time"] a_tokens_per_sec = llama_results["tokens_per_second"] if a_tokens_per_sec > 0: speedup = d_tokens_per_sec / a_tokens_per_sec else: speedup = 0 comparison_html = f"""

⚡ Performance Comparison

Metric D2F-LLaDA-Instruct-8B LLaMA3-Instruct-8B Difference
Total tokens {d_tokens} {a_tokens} -
Generation time {d_time:.2f}s {a_time:.2f}s {"D2F-LLaDA is " + f"{(a_time/d_time):.1f}x faster" if d_time > 0 and d_time < a_time else "LLaMA3 is " + f"{(d_time/a_time):.1f}x faster"}
Tokens per second {d_tokens_per_sec:.2f} {a_tokens_per_sec:.2f} {"D2F-LLaDA is " + f"{speedup:.1f}x faster" if speedup > 1 else "LLaMA3 is " + f"{(1/speedup if speedup > 0 else 0):.1f}x faster"}
""" return comparison_html def create_stats_html(model_name, results): stats_html = f"""

✓ {model_name} Generation Complete

""" return stats_html # --- Main Interface --- if __name__ == "__main__": os.environ["CUDA_VISIBLE_DEVICES"] = "3,4" torch.cuda.empty_cache() d2f_config = { "pretrained_path": "GSAI-ML/LLaDA-8B-Instruct", "lora_path": "SJTU-Deng-Lab/D2F_LLaDA_Instruct_8B_Lora", "device": "cuda:0", "dtype": "bfloat16", "max_length": 4096, "temperature": 0.0, "top_p": None, "top_k": None, "mask_token_id": 126336, "sampling_strategy": "default", } llama_config = { "model_id": "meta-llama/Llama-3.1-8B-Instruct", "device": "cuda:1", } set_seed(42) d2f_engine = D2FInference(**d2f_config) llama_engine = LlamaInference(**llama_config) with gr.Blocks(css=D2FInference.CSS, theme=gr.themes.Soft()) as demo: gr.Markdown("# 🚀 D2F-LLaDA vs LLaMA3: Speed Comparison") with gr.Row(): with gr.Column(scale=1): prompt_input = gr.Textbox( label="Enter your question", placeholder="Example: Natalia sold clips to...", lines=5 ) generate_button = gr.Button("🚀 Run Speed Comparison", variant="primary") with gr.Accordion("⚙️ D2F-LLaDA Parameter Settings", open=True): with gr.Row(): max_new_tokens_slider = gr.Slider( minimum=64, maximum=2048, value=1024, step=64, label="Max Tokens to Generate" ) block_size_slider = gr.Slider( minimum=16, maximum=128, value=32, step=16, label="Block Size" ) with gr.Row(): block_add_thresh_slider = gr.Slider( minimum=0.0, maximum=1.0, value=0.1, step=0.05, label="Block Add Threshold" ) decoded_token_thresh_slider = gr.Slider( minimum=0.0, maximum=1.0, value=0.5, step=0.05, label="Decoding Completion Threshold" ) skip_thresh_slider = gr.Slider( minimum=0.0, maximum=1.0, value=0.9, step=0.01, label="Skip Threshold" ) comparison_output = gr.HTML(label="Performance Comparison", elem_id="comparison-container") with gr.Row(): with gr.Column(scale=1): gr.HTML("
✨ D2F-LLaDA-Instruct-8B (Parallel Decoding)
") d2f_output = gr.Textbox( label="D2F-LLaDA Output", interactive=False, elem_classes=["output-textbox"] ) d2f_status = gr.HTML(label="D2F-LLaDA Stats") with gr.Column(scale=1): gr.HTML("
🔄 LLaMA3-Instruct-8B (Standard)
") llama_output = gr.Textbox( label="LLaMA3 Output", interactive=False, elem_classes=["output-textbox"] ) llama_status = gr.HTML(label="LLaMA3 Stats") gr.Examples( examples=[ ["Solve the equation x² - 6x + 8 = 0. First, explain what a quadratic equation is and why it can have up to two solutions. Then solve this equation using three different methods: factoring, completing the square, and the quadratic formula. For each method, explain the mathematical reasoning behind it, show all steps in detail, and discuss when this particular method is most useful. Finally, verify your solutions by substituting them back into the original equation.", 1024, 32, 0.1, 0.55, 0.9], ["A circular swimming pool has a diameter of 8 meters. Calculate the pool's circumference and area. First, explain the relationship between diameter, radius, circumference, and area of a circle, including the role of π in these formulas. Then perform the calculations using π ≈ 3.14159. Next, estimate how much water (in cubic meters) would be needed to fill this pool if it has a uniform depth of 1.5 meters. Finally, calculate how much it would cost to fill this pool if water costs $2.50 per cubic meter. Show all steps and include appropriate units in your answer.", 1024, 32, 0.1, 0.5, 0.9], ["A movie theater offers a loyalty card that costs $15 and gives a 15% discount on all tickets. If a regular movie ticket costs $10, how many tickets would you need to buy to make the loyalty card worthwhile? First, explain the concept of a break-even point. Then set up an equation to find when the total cost with the card equals the total cost without the card. Solve this equation step by step, showing all your work. Finally, interpret your answer in the context of the problem.", 1024, 32, 0.1, 0.5, 0.9], ], inputs=[ prompt_input, max_new_tokens_slider, block_size_slider, block_add_thresh_slider, decoded_token_thresh_slider, skip_thresh_slider ], label="Examples (Math Problems)" ) def run_models_streaming( prompt_text, max_new_tokens, block_size, block_add_threshold, decoded_token_threshold, skip_threshold ): torch.cuda.empty_cache() d2f_generator = d2f_engine.stream( prompt_text=prompt_text, max_new_tokens=max_new_tokens, block_size=block_size, block_add_threshold=block_add_threshold, decoded_token_threshold=decoded_token_threshold, skip_threshold=skip_threshold ) llama_generator = llama_engine.stream( prompt_text=prompt_text, max_new_tokens=max_new_tokens ) d2f_text = "" llama_text = "" d2f_stats = None llama_stats = None yield d2f_text, llama_text, "", "", "" d2f_done = False llama_done = False while not (d2f_done and llama_done): if not d2f_done: try: new_d2f_text, new_d2f_stats = next(d2f_generator) d2f_text = new_d2f_text if new_d2f_stats is not None: d2f_stats = new_d2f_stats d2f_done = True except StopIteration: d2f_done = True if not llama_done: try: new_llama_text, new_llama_stats = next(llama_generator) llama_text = new_llama_text if new_llama_stats is not None: llama_stats = new_llama_stats llama_done = True except StopIteration: llama_done = True d2f_status_html = create_stats_html("D2F-LLaDA", d2f_stats) if d2f_stats else "" llama_status_html = create_stats_html("LLaMA3", llama_stats) if llama_stats else "" comparison = "" if d2f_done and llama_done and d2f_stats and llama_stats: comparison = create_comparison_html(d2f_stats, llama_stats) yield d2f_text, llama_text, d2f_status_html, llama_status_html, comparison # MODIFICATION: Removed the _js parameter from here generate_button.click( fn=run_models_streaming, inputs=[ prompt_input, max_new_tokens_slider, block_size_slider, block_add_thresh_slider, decoded_token_thresh_slider, skip_thresh_slider ], outputs=[ d2f_output, llama_output, d2f_status, llama_status, comparison_output ] ) # MODIFICATION: Added a hidden HTML component with a script for auto-scrolling # This method is compatible with older Gradio versions. gr.HTML( """ """, visible=False ) demo.queue().launch(share=True)