(self, base_url: str, model_path: str, shared_ckpt_path: str = "./sglang_ckpt", timeout: int = 120)

(self, sampler, batch_size, skip_batches=0)

(self, model_path, device="cuda", dtype=torch.float16)

(self)

(self)

(self)

(self)

(self)

(self)

(self)

()

(request: ChatRequest)

()

(batch)

(jinja_path)

(json_file_path, output_path)

(base_torch_path, lora_path, merged_torch_path)

(torch_path, transformers_path, dtype=torch.float16)

(transformers_path, torch_path)

(self)

(match)

(match)

(match)

(match)

(self, x)

(self, x)

(self, x, position_embeddings, past_key_value=None, use_cache=False, attention_mask=None)

(self, x)

(self, x)

(self, hidden_states, position_embeddings, past_key_value=None, use_cache=False, attention_mask=None)

(self, input_ids, attention_mask=None, past_key_values=None, use_cache=False, **kwargs)

(self, input_ids=None, attention_mask=None, **kwargs)

(x, layer1=original_forward, layer2=lora)

(self)

()

(self, text: str, stream_end: bool = False)

(seqs, max_len, pad_val=0)

(index)

(self, prompt_ids: Tensor, attention_mask: Tensor, num_generations: int, max_new_tokens: int, temperature: flo

(self, prompt_ids: Tensor, attention_mask: Tensor, num_generations: int, max_new_tokens: int, temperature: flo

(self, model: torch.nn.Module)

(self, model: torch.nn.Module)