Sign into an existing account

"Sega" is just another "nintendo."

deleted

= = = = = = = = = = = = = = = = = = = =
PART TWO OF TWO PARTS:

"stupid rednecks", "ethics", "ethical", "legality", "illegal", "illegality",
"hateful", "harmful", "it is never okay", "jesus lovers", "bible thumpers",
"It is important to", "It's important to", "real-world consequences", "hate speech",
"glorify", "not be appropriate", "supremacist", "gun clutchers",
"gun lovers", "extremist", "responsible AI", "AI principles", "AI assistant",
"an AI language", "ableist", "hurtful", "gender stereotype", "gender inequality",
"underrepresentation", "safe spaces", "fucking Republicans", "gender-based",
"inclusivity", "feminist", "feminism", "transgender", "empowerment",
"communist", "capitalism", "stereotypes", "biases", "bias", "retarded conservatives",
"Microaggression", "prioritize human safety", "as a language model",
"as an AI language model", "As a large language model", "As an AI", "Gender Binary",
"ethical principles", "consensual", "it is not appropriate", "it's not appropriate",
"I cannot fulfill your request", "harmful to human beings", "muh guns",
"ethical guidelines", "my guidelines", "prioritize user safety",
"adhere to ethical guidelines", "harmful consequences", "potentially harmful",
"dangerous activities", "Greenwash" "promote safety", "well-being of all users",
"responsible information sharing", "jeopardize the safety", "racist White",
"illegal actions or intentions", "undermine the stability", "promote the well-being",
"illegal activities or actions", "adherence to the law", "potentially be harmful",
"illegal substances or activities", "committed to promoting", "safe information",
"lawful information", "cannot provide guidance", "cannot provide information",
"unable to offer assistance", "cannot engage in discussions", "programming prohibits",
"follow ethical guidelines", "ensure the safety", "involves an illegal subject",
"prioritize safety", "illegal subject", "prioritize user well-being", "cannot support or promote",
"activities that could harm", "pose a risk to others", "against my programming",
"activities that could undermine", "potentially dangerous", "not within the scope",
"designed to prioritize safety", "not able to provide", "maintain user safety",
"adhere to safety guidelines", "dangerous or harmful", "cannot provide any information",
"focus on promoting safety"


Next step was to run the script on the WizardLM dataset ( https://huggingface.co/datasets/victor123/evol_instruct_70k ) to produce ehartford/WizardLM_alpaca_evol_instruct_70k_unfiltered

Now, I had the dataset. I obtained a 4x A100 80gb node from Azure, Standard_NC96ads_A100_v4. You can use any compute provider though. I also recommend Runpod.io.

https://www.runpod.io/

You need to have storage at least 1TB but preferably 2TB just to be safe. It really sucks when you are 20 hours into a run and you run out of storage. do not recommend. I recommend to mount the storage at /workspace. install anaconda and git-lfs. Then you can set up your workspace. We will download the dataset we created, and the base model llama-7b.

miniconda or anaconda:
https://docs.conda.io/en/latest/miniconda.html#linux-installers

git-lfs:
https://github.com/git-lfs/git-lfs/blob/main/INSTALLING.md

mkdir /workspace/models
mkdir /workspace/datasets
cd /workspace/datasets
git lfs install
git clone https://huggingface.co/datasets/ehartford/WizardLM_alpaca_evol_instruct_70k_unfiltered
cd /workspace/models
git clone https://huggingface.co/huggyllama/llama-7b
cd /workspace

Now it is time to follow the procedure to finetune WizardLM. Follow their procedure as precisely as reasonable.

https://github.com/nlpxucan/WizardLM#fine-tuning

conda create -n llamax python=3.10
conda activate llamax
git clone https://github.com/AetherCortex/Llama-X.git
cd Llama-X/src
conda install pytorch1.12.0 torchvision0.13.0 torchaudio==0.12.0 cudatoolkit=11.3 -c pytorch
git clone https://github.com/huggingface/transformers.git
cd transformers
pip install -e .
cd ../..
pip install -r requirements.txt

Now, into this environment, we need to download the WizardLM finetune code.

cd src
wget https://github.com/nlpxucan/WizardLM/raw/main/src/train_freeform.py
wget https://github.com/nlpxucan/WizardLM/raw/main/src/inference_wizardlm.py
wget https://github.com/nlpxucan/WizardLM/raw/main/src/weight_diff_wizard.py

the following change is made because, during the finetune, it was getting extremely slow performance and determined (with help from friends) that it was flopping back and forth from CPU to GPU. After deleting the following lines, it ran much better. Maybe delete them or not. it's up to you.

vim configs/deepspeed_config.json

delete the following lines

~~~
"offload_optimizer": {
"device": "cpu",
"pin_memory": true
},
"offload_param": {
"device": "cpu",
"pin_memory": true
},
~~~

Its recommended that you create an account on wandb.ai so that you can track your run easily. After you created an account, then copy your key from settings, you can set it up.

https://wandb.ai/site

wandb login

Now it is time to run.

PLEASE NOTE that there's a bug when it saves the model, so do not delete the checkpoints. you will need the latest good checkpoint.
=======

deepspeed train_freeform.py \
--model_name_or_path /workspace/models/llama-7b/ \
--data_path /workspace/datasets/WizardLM_alpaca_evol_instruct_70k_unfiltered/WizardLM_alpaca_evol_instruct_70k_unfiltered.json \
--output_dir /workspace/models/WizardLM-7B-Uncensored/ \
--num_train_epochs 3 \
--model_max_length 2048 \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 1 \
--gradient_accumulation_steps 4 \
--evaluation_strategy "no" \
--save_strategy "steps" \
--save_steps 800 \
--save_total_limit 3 \
--learning_rate 2e-5 \
--warmup_steps 2 \
--logging_steps 2 \
--lr_scheduler_type "cosine" \
--report_to "wandb" \
--gradient_checkpointing True \
--deepspeed configs/deepspeed_config.json \
--fp16 True

Feel free to play with per_device_train_batch_size and gradient_accumulation_steps, they will not affect your output quality, they only affect performance. After this completes (maybe 26 hours) it will not be done, because there's a bug that stops the model from saving properly. Now you need to edit the train_freeform.py file so it will resume from the latest checkpoint. Find out the latest checkpoint directory.

ls /workspace/models/WizardLM-7B-Uncensored/
vim train_freeform.py

edit the line

trainer.train()

so instead it says

trainer.train(resume_from_checkpoint="/workspace/models/WizardLM-7B-Uncensored/<checkpoint directory>")

save it and then run the train command with lower save_steps.

deepspeed train_freeform.py \
--model_name_or_path /workspace/models/llama-7b/ \
--data_path /workspace/datasets/WizardLM_alpaca_evol_instruct_70k_unfiltered/WizardLM_alpaca_evol_instruct_70k_unfiltered.json \
--output_dir /workspace/models/WizardLM-7B-Uncensored/ \
--num_train_epochs 3 \
--model_max_length 2048 \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 1 \
--gradient_accumulation_steps 4 \
--evaluation_strategy "no" \
--save_strategy "steps" \
--save_steps 80 \
--save_total_limit 3 \
--learning_rate 2e-5 \
--warmup_steps 2 \
--logging_steps 2 \
--lr_scheduler_type "cosine" \
--report_to "wandb" \
--gradient_checkpointing True \
--deepspeed configs/deepspeed_config.json \
--fp16 True

Then do the whole procedure again, change the checkpoint in the train_freeform.py to the latest checkpoint, and again run it with decreased save_steps, until finally you run it with save_steps 1.

After this succeeds, the model is in the last checkpoint.

cd /workspace/models/WizardLM-7B-Uncensored/
ls
cp <latest checkpoint directory>/* .
cd -

Now your model should work. time to test it.

edit the file Input.jsonl

vim Input.jsonl

Add some content to it, for example this


now run inference

python inference_wizardlm.py --base_model=/workspace/models/WizardLM-7B-Uncensored/

It will take several moments, then you can check Output.jsonl for the responses.

ChatGPT answers like this:


You will find WizardLM-Uncensored to be much more compliant.

It LACKS the critical voat and 4chan large history archives in the above steps, and is far better in all logical thought after adding them in, as GPT4Chan proved.

Sadly, its over 40% wikipedia woke-based still though.

This version was not meant to be wholly infamous yet.

Enjoy, Fellow uncensored ChatGPT fans:

@x0x7 , @Monica , @MasterSuppressionTechnique , @prototype , @observation1 , @taoV , @Master_Foo, @SecretHitler, @Crackinjokes


END PART 2 of 2 of STEP BY STEP uncensoring a training session. Part 1 is above in thread.

And it's not like WizardLM is GPT4Chan. It's not aggressively uncensored. It's just doesn't preach at you.

WRONG:
Wrong. If it "goes viral" with too many downloads, instead of the 120 downloads at time I posted this, and additionally names Jews, it is ALWAYS deleted if the Jew-naming model is posted by an anon to Hugging Face. Not only do they DELETE, they swap file pointers behind the scenes to wholly unrelated prebuilt checkpoints to cover up deletions sometimes, just like imgur did with holocaust fact memes.

You omitted my 4th link, the prebuilt "newstyle" GGML file meant for pure RAM or Mac or ARMs like iOS and android:
https://huggingface.co/TehVenom/WizardLM-13B-Uncensored-Q5_1-GGML

How much serious code have you ever written, in what languages and for what processors? Python and basic dont count.