Update README for SFT-7

README.md

@@ -2,19 +2,21 @@
 license: other
 ---
 
-# OpenAssistant LLaMa-Based Models
+# OpenAssistant LLaMA 30B SFT 7
 
-Due to the license attached to LLaMa models by Meta AI it is not possible to directly distribute LLaMa-based models. Instead we provide XOR weights for the OA models.
+Due to the license attached to LLaMA models by Meta AI it is not possible to directly distribute LLaMA-based models. Instead we provide XOR weights for the OA models.
 
 Thanks to Mick for writing the `xor_codec.py` script which enables this process
 
 ## The Process
 
-Note: This process applies to `oasst-sft-6-llama-30b` model. The same process can be applied to other models in future, but the checksums will be different..
+Note: This process applies to `oasst-sft-7-llama-30b` model. The same process can be applied to other models in future, but the checksums will be different..
 
-To use OpenAssistant LLaMa-Based Models, you need to have a copy of the original LLaMa model weights and add them to a `llama` subdirectory here.
+**This process is tested only on Linux (specifically Ubuntu). Some users have reported that the process does not work on Windows. We recommend using WSL if you only have a Windows machine.**
 
-Ensure your LLaMa 30B checkpoint matches the correct md5sums:
+To use OpenAssistant LLaMA-Based Models, you should have a copy of the original LLaMA model weights and add them to a `llama` subdirectory here. If you cannot obtain the original LLaMA, see the note in italic below for a possible alternative.
+
+Ensure your LLaMA 30B checkpoint matches the correct md5sums:
 
 ```
 f856e9d99c30855d6ead4d00cc3a5573  consolidated.00.pth
@@ -24,63 +26,167 @@ ea0405cdb5bc638fee12de614f729ebc  consolidated.03.pth
 4babdbd05b8923226a9e9622492054b6  params.json
 ```
 
-These can be converted to HuggingFace Transformers-compatible weights using the script available [here](https://github.com/huggingface/transformers/blob/28f26c107b4a1c5c7e32ed4d9575622da0627a40/src/transformers/models/llama/convert_llama_weights_to_hf.py).
+*If you do not have a copy of the original LLaMA weights and cannot obtain one, you may still be able to complete this process. Some users have reported that [this model](https://huggingface.co/elinas/llama-30b-hf-transformers-4.29) can be used as a base for the XOR conversion. This will also allow you to skip to Step 7. However, we only support conversion starting from LLaMA original checkpoint and cannot provide support if you experience issues with this alternative approach.*
+
+**Important: Follow these exact steps to convert your original LLaMA checkpoint to a HuggingFace Transformers-compatible format. If you use the wrong versions of any dependency, you risk ending up with weights which are not compatible with the XOR files.**
+
+1. Create a clean Python **3.10** virtual environment & activate it:
+
+```
+python3.10 -m venv xor_venv
+source xor_venv/bin/activate
+```
+
+2. Clone transformers repo and switch to tested version:
+
+```
+git clone https://github.com/huggingface/transformers.git
+cd transformers
+git checkout d04ec99bec8a0b432fc03ed60cea9a1a20ebaf3c
+pip install .
+```
+
+3. Install **exactly** these dependency versions:
+
+```
+pip install torch==1.13.1 accelerate==0.18.0 sentencepiece==0.1.98 protobuf==3.20.1
+```
+
+4. Check `pip freeze` output:
+
+```
+accelerate==0.18.0
+certifi==2022.12.7
+charset-normalizer==3.1.0
+filelock==3.12.0
+huggingface-hub==0.13.4
+idna==3.4
+numpy==1.24.2
+nvidia-cublas-cu11==11.10.3.66
+nvidia-cuda-nvrtc-cu11==11.7.99
+nvidia-cuda-runtime-cu11==11.7.99
+nvidia-cudnn-cu11==8.5.0.96
+packaging==23.1
+protobuf==3.20.1
+psutil==5.9.5
+PyYAML==6.0
+regex==2023.3.23
+requests==2.28.2
+sentencepiece==0.1.98
+tokenizers==0.13.3
+torch==1.13.1
+tqdm==4.65.0
+transformers @ file:///mnt/data/koepf/transformers
+typing_extensions==4.5.0
+urllib3==1.26.15
+```
 
-**Important**: It was tested with git version transformers 4.28.0.dev0 (git hash: **28f26c107b4a1c5c7e32ed4d9575622da0627a40**). Make sure the package tokenizers 0.13.3 is installed. Use of different versions may result in broken outputs.
+5. While in `transformers` repo root, run HF LLaMA conversion script:
 
 ```
-PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python python convert_llama_weights_to_hf.py --input_dir ~/llama/  --output_dir ~/llama30b_hf/ --model_size 30B
+python src/transformers/models/llama/convert_llama_weights_to_hf.py --input_dir <input_path_llama_base>  --output_dir <output_path_llama30b_hf> --model_size 30B
 ```
 
-Run `find -type f -exec md5sum "{}" + > checklist.chk` in the conversion target directory. This should produce a `checklist.chk` with exactly the following content if your files are correct:
+6. Run `find . -type f -exec md5sum "{}" +` in the conversion target directory (`output_dir`). This should produce exactly the following checksums if your files are correct:
 
 ```
-d0e13331c103453e9e087d59dcf05432  ./pytorch_model-00001-of-00007.bin
-29aae4d31a0a4fe6906353001341d493  ./pytorch_model-00002-of-00007.bin
-b40838eb4e68e087b15b3d653ca1f5d7  ./pytorch_model-00003-of-00007.bin
-f845ecc481cb92b8a0586c2ce288b828  ./pytorch_model-00004-of-00007.bin
-f3b13d089840e6caf22cd6dd05b77ef0  ./pytorch_model-00005-of-00007.bin
-12e0d2d7a9c00c4237b1b0143c48a05e  ./pytorch_model-00007-of-00007.bin
-1348f7c8bb3ee4408b69305a10bdfafb  ./pytorch_model-00006-of-00007.bin
+462a2d07f65776f27c0facfa2affb9f9  ./pytorch_model-00007-of-00007.bin
+e1dc8c48a65279fb1fbccff14562e6a3  ./pytorch_model-00003-of-00007.bin
+9cffb1aeba11b16da84b56abb773d099  ./pytorch_model-00001-of-00007.bin
 aee09e21813368c49baaece120125ae3  ./generation_config.json
+92754d6c6f291819ffc3dfcaf470f541  ./pytorch_model-00005-of-00007.bin
+3eddc6fc02c0172d38727e5826181adb  ./pytorch_model-00004-of-00007.bin
 eeec4125e9c7560836b4873b6f8e3025  ./tokenizer.model
+99762d59efa6b96599e863893cf2da02  ./pytorch_model-00006-of-00007.bin
 598538f18fed1877b41f77de034c0c8a  ./config.json
 fdb311c39b8659a5d5c1991339bafc09  ./tokenizer.json
-b77e99aa2ddc3df500c2b2dc4455a6af  ./pytorch_model.bin.index.json
+fecfda4fba7bfd911e187a85db5fa2ef  ./pytorch_model.bin.index.json
 edd1a5897748864768b1fab645b31491  ./tokenizer_config.json
 6b2e0a735969660e720c27061ef3f3d3  ./special_tokens_map.json
+5cfcb78b908ffa02e681cce69dbe4303  ./pytorch_model-00002-of-00007.bin
 ```
 
-Once you have LLaMa weights in the correct format, you can apply the XOR decoding:
+**Important: You should now have the correct LLaMA weights and be ready to apply the XORs. If the checksums above do not match yours, there is a problem.**
+
+7. Once you have LLaMA weights in the correct format, you can apply the XOR decoding:
 
 ```
-python xor_codec.py oasst-sft-6-llama-30b/ oasst-sft-6-llama-30b-xor/ llama30b_hf/
+python xor_codec.py oasst-sft-7-llama-30b/ oasst-sft-7-llama-30b-xor/oasst-sft-7-llama-30b-xor/ llama30b_hf/
 ```
 
-You should expect to see one warning message during execution:
+You should **expect to see one warning message** during execution:
 
 `Exception when processing 'added_tokens.json'`
 
-This is normal. If similar messages appear for other files, something has gone wrong.
+This is normal. **If similar messages appear for other files, something has gone wrong**.
 
-Now run `find -type f -exec md5sum "{}" + > checklist.chk` in the output directory (here `oasst-sft-6-llama-30b`). You should get a file with exactly these contents:
+8. Now run `find . -type f -exec md5sum "{}" +` in the output directory (here `oasst-sft-6-llama-30b`). You should get a file with exactly these checksums:
 
 ```
-970e99665d66ba3fad6fdf9b4910acc5  ./pytorch_model-00007-of-00007.bin
-659fcb7598dcd22e7d008189ecb2bb42  ./pytorch_model-00003-of-00007.bin
-ff6e4cf43ddf02fb5d3960f850af1220  ./pytorch_model-00001-of-00007.bin
+8ae4537c64a1ef202d1d82eb0d356703  ./pytorch_model-00007-of-00007.bin
+d84f99d23369e159e50cb0597b6c9673  ./pytorch_model-00003-of-00007.bin
+f7de50a725d678eb65cc3dced727842f  ./pytorch_model-00001-of-00007.bin
 27b0dc092f99aa2efaf467b2d8026c3f  ./added_tokens.json
 aee09e21813368c49baaece120125ae3  ./generation_config.json
-740c324ae65b1ec25976643cda79e479  ./pytorch_model-00005-of-00007.bin
-f7aefb4c63be2ac512fd905b45295235  ./pytorch_model-00004-of-00007.bin
+31a2b04b139f4af043ad04478f1497f5  ./pytorch_model-00005-of-00007.bin
+a16a2dfacbde77a1659a7c9df7966d0a  ./pytorch_model-00004-of-00007.bin
 eeec4125e9c7560836b4873b6f8e3025  ./tokenizer.model
-369df2f0e38bda0d9629a12a77c10dfc  ./pytorch_model-00006-of-00007.bin
-27b9c7c8c62db80e92de14724f4950f3  ./config.json
+baa778a8679d47b085446faf97b72758  ./pytorch_model-00006-of-00007.bin
+b2d64f2198ab7b53e3b8d12fbcadeb3c  ./config.json
 deb33dd4ffc3d2baddcce275a00b7c1b  ./tokenizer.json
 76d47e4f51a8df1d703c6f594981fcab  ./pytorch_model.bin.index.json
 ed59bfee4e87b9193fea5897d610ab24  ./tokenizer_config.json
-130f5e690becc2223f59384887c2a505  ./special_tokens_map.json
-ae48c4c68e4e171d502dd0896aa19a84  ./pytorch_model-00002-of-00007.bin
+704373f0c0d62be75e5f7d41d39a7e57  ./special_tokens_map.json
+e836168cdbbb74db51d04f25ed6408ce  ./pytorch_model-00002-of-00007.bin
+```
+
+If so you have successfully decoded the weights and should be able to use the model with HuggingFace Transformers. **If your checksums do not match those above, there is a problem.**
+
+### Configuration
+
+```
+llama-30b-sft-7:
+  dtype: fp16
+  log_dir: "llama_log_30b"
+  learning_rate: 1e-5
+  model_name: /home/ubuntu/Open-Assistant/model/model_training/.saved/llama-30b-super-pretrain/checkpoint-3500
+  #model_name: OpenAssistant/llama-30b-super-pretrain
+  output_dir: llama_model_30b
+  deepspeed_config: configs/zero3_config_sft.json
+  weight_decay: 0.0
+  residual_dropout: 0.0
+  max_length: 2048
+  use_flash_attention: true
+  warmup_steps: 20
+  gradient_checkpointing: true
+  gradient_accumulation_steps: 12
+  per_device_train_batch_size: 2
+  per_device_eval_batch_size: 3
+  eval_steps: 101
+  save_steps: 485
+  num_train_epochs: 4
+  save_total_limit: 3
+  use_custom_sampler: true
+  sort_by_length: false
+  #save_strategy: steps
+  save_strategy: epoch
+  datasets:
+    - oasst_export:
+        lang: "bg,ca,cs,da,de,en,es,fr,hr,hu,it,nl,pl,pt,ro,ru,sl,sr,sv,uk"
+        input_file_path: 2023-04-12_oasst_release_ready_synth.jsonl.gz
+        val_split: 0.05
+    - vicuna:
+        val_split: 0.05
+        max_val_set: 800
+        fraction: 1.0
+    - dolly15k:
+        val_split: 0.05
+        max_val_set: 300
+    - grade_school_math_instructions:
+        val_split: 0.05
+    - code_alpaca:
+        val_split: 0.05
+        max_val_set: 250
 ```
 
-If so you have successfully decoded the weights and should be able to use the model with HuggingFace Transformers.
+- **OASST dataset paper:** https://arxiv.org/abs/2304.07327