char-fork/llama.cpp

git clone https://git.t4t.associates/char-fork/llama.cpp

Johannes Gäßlergithub: mention --log-file in issue templates (#23277)a8078675a

char/aliceai-t5-moe
brancheschar/aliceai-t5-moeall refs →
4.5 KiB117 linesraw
1name: Bug (model use)
2description: Something goes wrong when running a model (crashes, garbled outputs, etc.).
3title: "Eval bug: "
4labels: ["bug-unconfirmed", "model evaluation"]
5body:
6  - type: markdown
7    attributes:
8      value: >
9        Thanks for taking the time to fill out this bug report!
10        This issue template is intended for bug reports where the model evaluation results
11        (i.e. the generated text) are incorrect or llama.cpp crashes during model evaluation.
12        If you encountered the issue while using an external UI (e.g. ollama),
13        please reproduce your issue using one of the examples/binaries in this repository.
14        The `llama-completion` binary can be used for simple and reproducible model inference.
15
16        Please fill out this template yourself, copypasting language model outputs is [strictly prohibited](https://github.com/ggml-org/llama.cpp/blob/master/CONTRIBUTING.md#ai-usage-policy).
17  - type: textarea
18    id: version
19    attributes:
20      label: Name and Version
21      description: Which version of our software are you running? (use `--version` to get a version string)
22      placeholder: |
23        $./llama-cli --version
24        version: 2999 (42b4109e)
25        built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu
26    validations:
27      required: true
28  - type: dropdown
29    id: operating-system
30    attributes:
31      label: Operating systems
32      description: Which operating systems do you know to be affected?
33      multiple: true
34      options:
35        - Linux
36        - Mac
37        - Windows
38        - BSD
39        - Other? (Please let us know in description)
40    validations:
41      required: true
42  - type: dropdown
43    id: backends
44    attributes:
45        label: GGML backends
46        description: Which GGML backends do you know to be affected?
47        options: [AMX, BLAS, CANN, CPU, CUDA, Hexagon, HIP, Metal, Musa, OpenCL, OpenVINO, RPC, SYCL, VirtGPU, Vulkan, WebGPU, zDNN, ZenDNN]
48        multiple: true
49    validations:
50      required: true
51  - type: textarea
52    id: hardware
53    attributes:
54      label: Hardware
55      description: Which CPUs/GPUs are you using?
56      placeholder: >
57        e.g. Ryzen 5950X + 2x RTX 4090
58    validations:
59      required: true
60  - type: textarea
61    id: model
62    attributes:
63      label: Models
64      description: >
65        Which model(s) at which quantization were you using when encountering the bug?
66        If you downloaded a GGUF file off of Huggingface, please provide a link.
67      placeholder: >
68        e.g. Meta LLaMA 3.1 Instruct 8b q4_K_M
69    validations:
70      required: false
71  - type: textarea
72    id: info
73    attributes:
74      label: Problem description & steps to reproduce
75      description: >
76        Please give us a summary of the problem and tell us how to reproduce it.
77        If you can narrow down the bug to specific hardware, compile flags, or command line arguments,
78        that information would be very much appreciated by us.
79
80        If possible, please try to reproduce the issue using `llama-completion` with `-fit off`.
81        If you can only reproduce the issue with `-fit on`, please provide logs both with and without `--verbose`.
82      placeholder: >
83        e.g. when I run llama-completion with `-fa on` I get garbled outputs for very long prompts.
84        With short prompts or `-fa off` it works correctly.
85        Here are the exact commands that I used: ...
86    validations:
87      required: true
88  - type: textarea
89    id: first_bad_commit
90    attributes:
91      label: First Bad Commit
92      description: >
93        If the bug was not present on an earlier version: when did it start appearing?
94        If possible, please do a git bisect and identify the exact commit that introduced the bug.
95    validations:
96      required: false
97  - type: textarea
98    id: logs
99    attributes:
100      label: Relevant log output
101      description: >
102          Please copy and paste any relevant log output, including the command that you entered and any generated text.
103          For very long logs (thousands of lines), please upload them as files instead; the `--log-file` CLI argument can be used for this purpose.
104          On Linux you can alternatively redirect the console output of any command into a file by appending ` > llama.log 2>&1` to your command.
105      value: |
106        <details>
107        <summary>Logs</summary>
108        <!-- Copy-pasted short logs go into the "console" area here -->
109
110        ```console
111
112        ```
113        </details>
114
115        <!-- Long logs that you upload as files go here, outside the "console" area -->
116    validations:
117      required: true