Closed Arcmoon-Hu closed 1 month ago
Could you turn on RUST_LOG=debug
and share its logs? Besides, we do support native Windows installation; you can find the instructions at https://tabby.tabbyml.com/docs/quick-start/installation/windows/, which might help you avoid dealing with WSL hassles.
I am having same issue on 0.13.
Jul 19 10:22:02 hre-gn0197 systemd[1]: Started TabbyML.
Jul 19 10:22:02 hre-gn0197 tabby[898068]: 2024-07-19T15:22:02.010550Z DEBUG tabby_common::config: crates/tabby-common/src/config.rs:35: Config file /root/.tabby/config.toml not found, apply default configuration
Jul 19 10:22:02 hre-gn0197 tabby[898068]: 2024-07-19T15:22:02.010857Z INFO tabby::services::model: crates/tabby/src/services/model/mod.rs:44: Loading model from local path /software/tabbyml/CodeLlama-13B
Jul 19 10:22:02 hre-gn0197 tabby[898068]: 2024-07-19T15:22:02.011107Z INFO tabby::services::model: crates/tabby/src/services/model/mod.rs:44: Loading model from local path /software/tabbyml/Mistral-7B
Jul 19 10:22:02 hre-gn0197 tabby[898068]: 2024-07-19T15:22:02.018012Z DEBUG reqwest::connect: /root/.cargo/registry/src/index.crates.io-6f17d22bba15001f/reqwest-0.12.4/src/connect.rs:497: starting new connection: https://raw.githubusercontent.com/
This is an air gapped server so no direct connection to the internet. I have configured crates proxy through Artifactory. How can we get this connection to not be attempted on startup?
Caused by:
0: error sending request for url (https://raw.githubusercontent.com/TabbyML/registry-tabby/main/models.json)
1: client error (Connect)
2: error:1416F086:SSL routines:tls_process_server_certificate:certificate verify failed:ssl/statem/statem_clnt.c:1915: (Hostname mismatch)
3: error:1416F086:SSL routines:tls_process_server_certificate:certificate verify failed:ssl/statem/statem_clnt.c:1915:
I already downloaded the models using tabby download
so unsure why this is being attempted?
[root@server~]# ls -ltr /software/tabbyml/CodeLlama-13B/
total 1
-rw-r--r-- 1 root root 254 Jul 19 09:50 tabby.json
drwxr-xr-x 2 root root 4096 Jul 19 09:56 ggml
[root@server~]# ls -ltr /software/tabbyml/Mistral-7B/
total 1
-rw-r--r-- 1 root root 694 Jul 19 09:58 tabby.json
drwxr-xr-x 2 root root 4096 Jul 19 10:00 ggml
Describe the bug Use wsl start server and follow docs, the command is
./tabby serve --model $LLM_MODEL_PATH --chat-model $LLM_MODEL_PATH
The log shows that the model exists and is loading.
But it consume long time and never changed.
Information about your version 0.13.1
Information about your GPU![image](https://github.com/user-attachments/assets/d78f8db2-6706-4b11-9982-e9ba28be78dd)