Mps mistral lora (#1292) [skip ci]

* Lora example for Mistral on MPS backend

* Add some MPS documentation

* Update examples/mistral/lora-mps.yml

Co-authored-by: NanoCode012 <[email protected]>

* Update examples/mistral/lora-mps.yml

Co-authored-by: NanoCode012 <[email protected]>

* Update README.md

---------

Co-authored-by: NanoCode012 <[email protected]>
Co-authored-by: Wing Lian <[email protected]>

.gitignore

@@ -167,3 +167,8 @@ cython_debug/
 # WandB
 # wandb creates a folder to store logs for training runs
 wandb
+
+# Runs
+lora-out/*
+qlora-out/*
+mlruns/*

README.md

@@ -99,7 +99,23 @@ Get started with Axolotl in just a few steps! This quickstart guide will walk yo
 
 **Requirements**: Python >=3.9 and Pytorch >=2.1.1.
 
-`pip3 install "axolotl[flash-attn,deepspeed] @ git+https://github.com/OpenAccess-AI-Collective/axolotl"`
+### For developers
+```bash
+git clone https://github.com/OpenAccess-AI-Collective/axolotl
+cd axolotl
+
+pip3 install packaging
+```
+
+General case:
+```
+pip3 install -e '.[flash-attn,deepspeed]'
+```
+
+Mac: see https://github.com/OpenAccess-AI-Collective/axolotl/blob/13199f678b9aab39e92961323bdbce3234ee4b2b/docs/mac.md
+```
+pip3 install -e '.'
+```
 
 ### Usage
 ```bash

docs/mac.md

@@ -0,0 +1,18 @@
+# Mac M series support
+
+Currently Axolotl on Mac is partially usable, many of the dependencies of Axolotl including Pytorch do not support MPS or have incomplete support.
+
+Current support:
+- [x] Support for all models
+- [x] Full training of models
+- [x] LoRA training
+- [x] Sample packing
+- [ ] FP16 and BF16 (awaiting AMP support for MPS in Pytorch)
+- [ ] Tri-dao's flash-attn (until it is supported use spd_attention as an alternative)
+- [ ] xformers
+- [ ] bitsandbytes (meaning no 4/8 bits loading and bnb optimizers)
+- [ ] qlora
+- [ ] DeepSpeed
+
+Untested:
+- FSDP

examples/mistral/lora-mps.yml

@@ -0,0 +1,79 @@
+base_model: mistralai/Mistral-7B-v0.1
+model_type: MistralForCausalLM
+tokenizer_type: LlamaTokenizer
+
+load_in_8bit: false
+load_in_4bit: false
+strict: false
+
+datasets:
+  - path: mhenrichsen/alpaca_2k_test
+    type: alpaca
+dataset_prepared_path: last_run_prepared
+val_set_size: 0
+output_dir: ./lora-out
+eval_sample_packing: false
+
+adapter: lora
+lora_model_dir:
+
+sequence_len: 4096
+sample_packing: true
+pad_to_sequence_len: true
+
+lora_r: 32
+lora_alpha: 16
+lora_dropout: 0.05
+lora_target_linear: true
+lora_fan_in_fan_out:
+lora_target_modules:
+  - gate_proj
+  - down_proj
+  - up_proj
+  - q_proj
+  - v_proj
+  - k_proj
+  - o_proj
+
+wandb_project:
+wandb_entity:
+wandb_watch:
+wandb_name:
+wandb_log_model:
+
+gradient_accumulation_steps: 8
+micro_batch_size: 1
+num_epochs: 2
+optimizer: adamw_torch
+lr_scheduler: cosine
+learning_rate: 0.0002
+
+train_on_inputs: false
+group_by_length: false
+bf16: auto
+fp16: false
+tf32: true
+
+gradient_checkpointing: true
+early_stopping_patience:
+resume_from_checkpoint:
+local_rank:
+logging_steps: 1
+xformers_attention:
+flash_attention: false
+sdp_attention: true
+
+loss_watchdog_threshold: 5.0
+loss_watchdog_patience: 3
+
+warmup_steps: 10
+evals_per_epoch: 4
+eval_table_size:
+eval_table_max_new_tokens: 128
+saves_per_epoch: 1
+debug:
+deepspeed:
+weight_decay: 0.0
+fsdp:
+fsdp_config:
+special_tokens: