Sketch a FAQ and add some issue templates.

.github/ISSUE_TEMPLATE/bug.yml

@@ -0,0 +1,83 @@
+name: Bug Report
+description: You found a bug.
+labels: ["bug", "triage"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Please first check the [FAQ](https://github.com/kyutai-labs/delayed-streams-modeling/blob/main/FAQ.md).
+  - type: dropdown
+    id: backend
+    attributes:
+      label: Backend impacted
+      description: Which backend is concerned with your bug report?
+      options:
+        - The PyTorch implementation
+        - The MLX implementation
+        - The Rust implementation
+        - Other / All
+      default: 0
+    validations:
+      required: true
+  - type: dropdown
+    id: os
+    attributes:
+      label: Operating system
+      description: What is your operating system?
+      options:
+        - Linux
+        - Mac OS X
+        - Windows (unsupported)
+      default: 0
+    validations:
+      required: true
+  - type: dropdown
+    id: hardware
+    attributes:
+      label: Hardware
+      description: What hardware are you using?
+      options:
+        - CPU
+        - GPU with CUDA
+        - Metal with MLX
+      default: 0
+    validations:
+      required: true
+  - type: textarea
+    id: description
+    attributes:
+      label: Description
+      description: Provide a detailed description of your bug.
+      placeholder: 
+      value: 
+    validations:
+      required: true
+  - type: textarea
+    id: more_info
+    attributes:
+      label: Extra information
+      description: Please provide any other relevant information, such as log extracts, code etc.
+      placeholder: 
+      value: 
+    validations:
+      required: true
+  - type: textarea
+    id: env
+    attributes:
+      label: Environment
+      description: Please provide any other relevant information, such as log extracts, code etc.
+      placeholder: 
+      value: |
+          Fill in the following information on your system.
+          - Operating system version:
+
+          If the backend impacted is PyTorch:
+          - Python version:
+          - PyTorch version:
+          - CUDA version (run `python -c 'import torch;  print(torch.version.cuda)'`):
+          - GPU model and memory:
+            
+          If the backend is MLX:
+          - Mac model:
+    validations:
+      required: true

.github/ISSUE_TEMPLATE/question.yml

@@ -0,0 +1,40 @@
+name: Question
+description: You have a question about the codebase, the paper, or the implementation.
+labels: ["question", "triage"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Please first check the [FAQ](https://github.com/kyutai-labs/delayed-streams-modeling/blob/main/FAQ.md).
+  - type: checkboxes
+    id: terms
+    attributes:
+      label: Due diligence
+      description: Have you searched the existing issues / FAQ / Google / asked ChatGPT?
+      options:
+        - label: I have done my due diligence in trying to find the answer myself.
+          required: true
+
+  - type: dropdown
+    id: backend
+    attributes:
+      label: Topic
+      description: What is your question about?
+      options:
+        - The paper
+        - The PyTorch implementation
+        - The MLX implementation
+        - The Rust implementation
+        - Other / All
+      default: 0
+    validations:
+      required: true
+  - type: textarea
+    id: question
+    attributes:
+      label: Question
+      description: What is your question?
+      placeholder: Your question. Please make sure this is directly related to our codebase. We will not provide support for installing PyTorch, CUDA, Rust etc.
+      value: 
+    validations:
+      required: true

.gitignore

@@ -192,3 +192,4 @@ cython_debug/
 #  refer to https://docs.cursor.com/context/ignore-files
 .cursorignore
 .cursorindexingignore
+out*.wav

FAQ.md

@@ -0,0 +1,30 @@
+# FAQ
+
+Here is the answer to a number of frequently asked questions.
+
+### Torch Compilation Errors
+
+With some PyTorch/triton versions, one might encounter compilation errors
+like the following:
+```
+  Traceback (most recent call last):
+  ...
+  File "site-packages/torch/_inductor/runtime/triton_heuristics.py", line 1153, in make_launcher
+    "launch_enter_hook": binary.__class__.launch_enter_hook,
+                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+torch._inductor.exc.InductorError: AttributeError: type object 'CompiledKernel' has no attribute 'launch_enter_hook'
+```
+
+If that's the case, you can disable torch compilation by setting the following
+environment variable.
+```bash
+export NO_TORCH_COMPILE=1
+```
+
+### Will you release training code?
+
+Some finetuning code can be found in the [kyutai-labs/moshi-finetune repo](https://github.com/kyutai-labs/moshi-finetune).
+This code has not been adapted to the Speech-To-Text and Text-To-Speech models
+yet, but it should be a good starting point.
+
+

README.md

@@ -305,6 +305,10 @@ If you have [uv](https://docs.astral.sh/uv/) installed, you can skip the install
 and just prefix the command above with `uvx --with moshi-mlx`.
 </details>
 
+## FAQ
+
+Checkout the [Frequently Asked Questions](FAQ.md) section before opening an issue.
+
 ## License
 
 The present code is provided under the MIT license for the Python parts, and Apache license for the Rust backend.