fix: qwen3-next to use fla causal-conv1d to support packing (#3437

* fix: qwen3-next to use fla causal-conv1d to support packing

* fix: causal import and update doc for v5

* fix: hard fail for packing without fla

examples/qwen3-next/README.md

@@ -6,30 +6,13 @@ This guide shows how to fine-tune it with Axolotl with multi-turn conversations
 
 ## Getting started
 
-1. Install Axolotl following the [installation guide](https://docs.axolotl.ai/docs/installation.html). You need to install from main as Qwen3-Next is only on nightly or use our latest [Docker images](https://docs.axolotl.ai/docs/docker.html).
+1. Install Axolotl following the [installation guide](https://docs.axolotl.ai/docs/installation.html).
 
-    Here is an example of how to install from main for pip:
-
-```bash
-# Ensure you have Pytorch installed (Pytorch 2.6.0 min)
-git clone https://github.com/axolotl-ai-cloud/axolotl.git
-cd axolotl
-
-pip3 install packaging==26.0 setuptools==75.8.0 wheel ninja
-pip3 install --no-build-isolation -e '.[flash-attn]'
-
-# Install CCE https://docs.axolotl.ai/docs/custom_integrations.html#cut-cross-entropy
-python scripts/cutcrossentropy_install.py | sh
-```
-
-2. Install Qwen3-Next transformers commit
-```bash
-pip3 uninstall -y transformers && pip3 install "git+https://github.com/huggingface/transformers.git@b9282355bea846b54ed850a066901496b19da654"
-```
+2. Install [Cut Cross Entropy](https://docs.axolotl.ai/docs/custom_integrations.html#cut-cross-entropy) to reduce training VRAM usage.
 
 3. Install FLA for improved performance
 ```bash
-pip3 uninstall -y causal-conv1d && pip3 install flash-linear-attention==0.3.2
+pip3 uninstall -y causal-conv1d && pip3 install flash-linear-attention==0.4.1
 ```
 
 4. Run the finetuning example:
@@ -38,7 +21,7 @@ pip3 uninstall -y causal-conv1d && pip3 install flash-linear-attention==0.3.2
 axolotl train examples/qwen3-next/qwen3-next-80b-a3b-qlora.yaml
 ```
 
-This config uses about 45.62 GiB VRAM.
+This config uses about ~47 GiB (no target experts) and ~71GiB (target experts) VRAM.
 
 Let us know how it goes. Happy finetuning! 🚀
 

examples/qwen3-next/qwen3-next-80b-a3b-qlora.yaml

@@ -9,6 +9,8 @@ plugins:
 load_in_8bit: false
 load_in_4bit: true
 
+quantize_moe_experts: true
+
 datasets:
   - path: fozziethebeat/alpaca_messages_2k_test
     type: chat_template
@@ -25,7 +27,7 @@ sample_packing: true
 
 lora_r: 16
 lora_alpha: 8
-lora_dropout: 0.05
+lora_dropout: 0
 lora_target_modules:
   - linear_attn.in_proj_ba
   - linear_attn.in_proj_qkvz
@@ -34,12 +36,19 @@ lora_target_modules:
   - shared_expert.down_proj
   - shared_expert.gate_proj
   - shared_expert_gate
-  - mlp.gate
   - q_proj
   - v_proj
   - k_proj
   - o_proj
 
+# lora_target_parameters:
+#   - mlp.experts.gate_up_proj
+#   - mlp.experts.down_proj
+
+lora_mlp_kernel: false
+lora_qkv_kernel: false
+lora_o_kernel: false
+
 wandb_project:
 wandb_entity:
 wandb_watch: