LMDrive

LMDrive copied to clipboard

The problem of model feature extraction from decoder

3

Hello, thank you for your excellent work. In memfuser.py, how are 2500 and 2501:2506 determined in the features queried from decoder? ```python traffic_feature = hs[:, :2500] traffic_light_state_feature = hs[:, 2500]...

pisces365

Dockerize the whole repo !

1

chrissunny94

Training time

1

Great work！ How much time does the training process require?

fcjian

@deepcs233 Hello, how can I remove the suddenly appearing pedestrians? In which folder should I modify to achieve this？

CoderXuans

Replace the vehicle

4

@deepcs233 I want to replace the ego vehicle with an unmanned boat during evaluation. Should I make changes from both aspects: the code and the simulation?

CoderXuans

只运行evaluate.py是否可以直接测试一个仅包含四路相机+激光点云的真实数据

1

您好，想请问下该怎么去执行推理过程，不测试指标，这样是可行的吗

leftdoor

"What’s the difference between the images: carlasim/carla:0.9.10.1 and the tarball CARLA_0.9.10.1.tar.gz？

1

docker pull carlasim/carla:0.9.10.1 failed, Error response from daemon: Get "https://registry-1.docker.io/v2/": net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers) is there any pratical methods to docker pull...

PonyAIjkz

LLM权重放在哪里呢？

1

你好： 我观察到yaml里有llm_model，这在您的说明中表示llm_model的ckpt，但是这/data/llava-v1.5-7b似乎是一个文件夹，从代码来看似乎需要tokenizer和llm的ckpt self.llm_tokenizer = LlamaTokenizer.from_pretrained(llm_model, use_fast=False, truncation_side="left") self.llm_model = LlamaForCausalLM.from_pretrained(llm_model, torch_dtype=torch.float16, low_cpu_mem_usage=True)

1476732094