feat: add doc for expert quantization, glm45 air example configs, and update readme for release (#3452) [skip ci]

* chore: rename without period

* feat: add glm45 air

* feat: add doc on expert quantization

* feat: update base readme with new changes

* chore: cleanup

* chore: cleanup

* chore: cleanup

* fix: disable quantize_moe_expert on merge per comment

* chore: add kernel info to optimizations doc

docs/expert_quantization.qmd

@@ -0,0 +1,66 @@
+---
+title: "MoE Expert Quantization"
+description: "Reduce VRAM usage when training MoE model adapters by quantizing expert weights on load"
+---
+
+Transformers v5 changed MoE expert layers from `nn.Linear` to fused `nn.Parameter` (3D+ tensors).
+This means `bitsandbytes` can no longer quantize them during model loading, resulting in all expert
+weights being loaded in full bf16 precision and causing massive VRAM usage.
+
+`quantize_moe_experts` solves this by quantizing expert weights during model loading.
+It intercepts the weight loading process, quantizes each expert tensor on the fly, and
+immediately frees the original bf16 tensor from VRAM. This dramatically reduces peak memory.
+For example, GLM-4.7-Flash QLoRA drops from ~127GiB to ~23GiB reserved memory.
+
+## Usage
+
+Enable expert quantization in your Axolotl config:
+
+```yaml
+quantize_moe_experts: true
+```
+
+This works with both 4-bit (QLoRA) and 8-bit (LoRA) quantization.
+
+### Expert LoRA targeting
+
+You can optionally apply LoRA adapters directly to expert weights using `lora_target_parameters`:
+
+```yaml
+lora_target_parameters:
+  - mlp.experts.gate_up_proj
+  - mlp.experts.down_proj
+  # - mlp.gate.weight  # router
+```
+
+::: {.callout-note}
+`lora_dropout` must be `0` when using `lora_target_parameters`.
+:::
+
+## Requirements
+
+- Requires (`adapter: lora` and `load_in_8bit: true`) or (`adapter: qlora` and `load_in_4bit: true`)
+- CUDA GPUs only (not tested with ROCm or other backends)
+- FSDP2 compatible for distributed training
+
+## Limitations
+
+- `cpu_ram_efficient_loading` hangs / takes long time with FSDP2 + QLoRA.
+- Total model parameter count may display incorrectly (trainable param count is correct).
+- FSDP LoRA (8-bit) may have a large initial VRAM spike at the first 1-2 steps, which then drops. QLoRA does not exhibit this.
+- FSDP2 may use more VRAM per GPU than single GPU training due to not all layers being properly sharded across ranks.
+- Model loading takes longer due to on-demand quantization, even on consecutive runs.
+- DeepSpeed has not been tested.
+
+## Implementation details
+
+The quantization is applied by patching transformers to intercept weight loading.
+When a 3D+ CUDA tensor with "expert" in its name is detected:
+
+- **4-bit mode:** Uses bitsandbytes NF4 parametrization (configurable via `bnb_4bit_quant_type`).
+- **8-bit mode:** Uses a custom row-wise int8 parametrization with bitsandbytes dequantization.
+
+The original bf16 tensor is freed immediately after quantization. Multiple sub-patches are applied to
+transformers, PEFT and accelerate FSDP2 to support these parametrized expert modules.
+
+For full implementation details, see [PR #3439](https://github.com/axolotl-ai-cloud/axolotl/pull/3439).