| { |
| "model_type": "helion-osc", |
| "architectures": ["HelionOSCForCausalLM"], |
| "vocab_size": 102400, |
| "hidden_size": 5120, |
| "num_hidden_layers": 48, |
| "num_attention_heads": 40, |
| "num_key_value_heads": 8, |
| "intermediate_size": 18432, |
| "hidden_act": "swiglu", |
| "max_position_embeddings": 262144, |
| "initializer_range": 0.02, |
| "rms_norm_eps": 1e-6, |
| "use_cache": true, |
| "pad_token_id": 0, |
| "bos_token_id": 1, |
| "eos_token_id": 2, |
| "tie_word_embeddings": false, |
| "rope_theta": 10000000.0, |
| "rope_scaling": { |
| "type": "linear", |
| "factor": 32.0 |
| }, |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "mlp_bias": false, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.40.0", |
| "model_version": "1.0", |
| "use_flash_attention": true, |
| "sliding_window": null, |
| "gradient_checkpointing": false, |
| "task_specific_params": { |
| "code_generation": { |
| "max_length": 4096, |
| "temperature": 0.7, |
| "top_p": 0.95, |
| "top_k": 50, |
| "do_sample": true, |
| "repetition_penalty": 1.05, |
| "length_penalty": 1.0 |
| }, |
| "mathematical_reasoning": { |
| "max_length": 2048, |
| "temperature": 0.3, |
| "top_p": 0.9, |
| "top_k": 40, |
| "do_sample": false, |
| "repetition_penalty": 1.0, |
| "length_penalty": 1.2 |
| }, |
| "code_completion": { |
| "max_length": 1024, |
| "temperature": 0.6, |
| "top_p": 0.92, |
| "top_k": 45, |
| "do_sample": true, |
| "repetition_penalty": 1.03, |
| "stop_sequences": ["\n\n", "```", "###"] |
| }, |
| "algorithm_design": { |
| "max_length": 3072, |
| "temperature": 0.5, |
| "top_p": 0.93, |
| "top_k": 50, |
| "do_sample": true, |
| "repetition_penalty": 1.08 |
| }, |
| "debugging": { |
| "max_length": 2048, |
| "temperature": 0.4, |
| "top_p": 0.88, |
| "do_sample": false, |
| "repetition_penalty": 1.0 |
| } |
| }, |
| "specialization": { |
| "domain": "coding_and_mathematics", |
| "primary_focus": "code_generation_with_mathematical_reasoning", |
| "verification_enabled": true, |
| "step_by_step_reasoning": true, |
| "languages_supported": [ |
| "python", |
| "javascript", |
| "typescript", |
| "java", |
| "c", |
| "cpp", |
| "csharp", |
| "go", |
| "rust", |
| "ruby", |
| "php", |
| "swift", |
| "kotlin", |
| "scala", |
| "r", |
| "sql", |
| "bash", |
| "shell" |
| ], |
| "features": [ |
| "code_generation", |
| "code_completion", |
| "bug_detection", |
| "bug_fixing", |
| "mathematical_reasoning", |
| "theorem_proving", |
| "algorithm_design", |
| "algorithm_optimization", |
| "code_refactoring", |
| "documentation_generation", |
| "test_generation", |
| "complexity_analysis" |
| ], |
| "mathematical_capabilities": [ |
| "arithmetic", |
| "algebra", |
| "calculus", |
| "discrete_mathematics", |
| "linear_algebra", |
| "probability", |
| "statistics", |
| "number_theory", |
| "graph_theory", |
| "combinatorics" |
| ] |
| }, |
| "training_config": { |
| "training_precision": "bf16", |
| "optimizer": "adamw", |
| "learning_rate": 2e-5, |
| "warmup_steps": 2000, |
| "weight_decay": 0.01, |
| "max_grad_norm": 1.0 |
| }, |
| "quantization_config": { |
| "quant_method": "bitsandbytes", |
| "load_in_8bit": false, |
| "load_in_4bit": false, |
| "bnb_4bit_compute_dtype": "bfloat16", |
| "bnb_4bit_use_double_quant": true, |
| "bnb_4bit_quant_type": "nf4" |
| }, |
| "generation_config": { |
| "temperature": 0.7, |
| "top_p": 0.95, |
| "top_k": 50, |
| "do_sample": true, |
| "max_new_tokens": 2048, |
| "min_new_tokens": 1, |
| "num_beams": 1, |
| "early_stopping": false, |
| "no_repeat_ngram_size": 3, |
| "encoder_no_repeat_ngram_size": 0, |
| "diversity_penalty": 0.0, |
| "repetition_penalty": 1.05, |
| "length_penalty": 1.0, |
| "exponential_decay_length_penalty": null |
| }, |
| "special_tokens": { |
| "pad_token": "<|pad|>", |
| "bos_token": "<|begin_of_text|>", |
| "eos_token": "<|end_of_text|>", |
| "unk_token": "<|unk|>", |
| "code_start_token": "<|code_start|>", |
| "code_end_token": "<|code_end|>", |
| "math_start_token": "<|math_start|>", |
| "math_end_token": "<|math_end|>", |
| "reasoning_start_token": "<|reasoning_start|>", |
| "reasoning_end_token": "<|reasoning_end|>", |
| "explanation_start_token": "<|explanation_start|>", |
| "explanation_end_token": "<|explanation_end|>" |
| }, |
| "supported_frameworks": [ |
| "pytorch", |
| "tensorflow", |
| "onnx", |
| "jax" |
| ], |
| "evaluation_metrics": { |
| "humaneval_pass_at_1": 0.852, |
| "humaneval_pass_at_10": 0.928, |
| "mbpp_pass_at_1": 0.795, |
| "mbpp_pass_at_10": 0.891, |
| "gsm8k_accuracy": 0.785, |
| "math_accuracy": 0.623, |
| "apps_accuracy": 0.412 |
| }, |
| "hardware_requirements": { |
| "minimum_vram_gb": 16, |
| "recommended_vram_gb": 24, |
| "minimum_ram_gb": 32, |
| "recommended_ram_gb": 64, |
| "cpu_cores": 8, |
| "gpu_support": true, |
| "multi_gpu_support": true, |
| "cpu_only_support": true |
| }, |
| "deployment_options": { |
| "inference_frameworks": [ |
| "vllm", |
| "text-generation-inference", |
| "ollama", |
| "llama.cpp" |
| ], |
| "optimization_support": [ |
| "quantization", |
| "pruning", |
| "distillation", |
| "tensorrt", |
| "onnx_runtime" |
| ] |
| } |
| } |
