Hugging Face's logo

TheBloke
/
Vicuna-7B-v1-3-SuperHOT-8K-GGML

Model card Files Community
Edit model card
TheBlokeAI

LmSys' Vicuna 7B v1.3 GGML

These files are GGML format model files for LmSys' Vicuna 7B v1.3.

These are SuperHOT GGMLs with an increased context length. SuperHOT is a new system that employs RoPE to expand context beyond what was originally possible for a model. It was discovered and developed by kaiokendev.

In order to use the increased context length, you can presently use:

Support is also expected to come to llama.cpp, however work is still being done to find the optimal implementation.

To use the increased context with KoboldCpp, simply use --contextsize to set the desired context, eg --contextsize 4096 or --contextsize 8192.

NOTE: Increased context length is an area seeing rapid developments and improvements. It is quite possible that these models may be superseded by new developments in the coming days. If that's the case, I will remove them, or update this README as appropriate.

Repositories available

Compatibility

These GGMLs will work with any llama.cpp-compatible GGML client that supports k-quants.

However the increased context length won't work without specific support. See the note in the introduction for details on using increased context.

Explanation of the new k-quant methods

The new methods available are:

  • GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
  • GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
  • GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
  • GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
  • GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
  • GGML_TYPE_Q8_K - "type-0" 8-bit quantization. Only used for quantizing intermediate results. The difference to the existing Q8_0 is that the block size is 256. All 2-6 bit dot products are implemented for this quantization type.

Refer to the Provided Files table below to see what files use which methods, and how.

Provided files

Name Quant method Bits Size Max RAM required Use case
vicuna-7b-v1.3-superhot-8k.ggmlv3.q2_K.bin q2_K 2 2.87 GB 5.37 GB New k-quant method. Uses GGML_TYPE_Q4_K for the attention.vw and feed_forward.w2 tensors, GGML_TYPE_Q2_K for the other tensors.
vicuna-7b-v1.3-superhot-8k.ggmlv3.q3_K_L.bin q3_K_L 3 3.60 GB 6.10 GB New k-quant method. Uses GGML_TYPE_Q5_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K
vicuna-7b-v1.3-superhot-8k.ggmlv3.q3_K_M.bin q3_K_M 3 3.28 GB 5.78 GB New k-quant method. Uses GGML_TYPE_Q4_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K
vicuna-7b-v1.3-superhot-8k.ggmlv3.q3_K_S.bin q3_K_S 3 2.95 GB 5.45 GB New k-quant method. Uses GGML_TYPE_Q3_K for all tensors
vicuna-7b-v1.3-superhot-8k.ggmlv3.q4_K_M.bin q4_K_M 4 4.08 GB 6.58 GB New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q4_K
vicuna-7b-v1.3-superhot-8k.ggmlv3.q4_K_S.bin q4_K_S 4 3.83 GB 6.33 GB New k-quant method. Uses GGML_TYPE_Q4_K for all tensors
vicuna-7b-v1.3-superhot-8k.ggmlv3.q5_K_M.bin q5_K_M 5 4.78 GB 7.28 GB New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q5_K
vicuna-7b-v1.3-superhot-8k.ggmlv3.q5_K_S.bin q5_K_S 5 4.65 GB 7.15 GB New k-quant method. Uses GGML_TYPE_Q5_K for all tensors
vicuna-7b-v1.3-superhot-8k.ggmlv3.q6_K.bin q6_K 6 5.53 GB 8.03 GB New k-quant method. Uses GGML_TYPE_Q8_K - 6-bit quantization - for all tensors

Note: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.

How to run in koboldcpp

On Linux I use the following command line to launch the KoboldCpp UI with OpenCL aceleration and a context size of 4096:

python ./koboldcpp.py --stream --unbantokens --threads 8 --usecublas --gpulayers 100 vicuna-7b-v1.3-superhot-8k.ggmlv3.q4_K_M.bin

Change --gpulayers 100 to the number of layers you want/are able to offload to the GPU. Remove it if you don't have GPU acceleration.

For OpenCL acceleration, change --usecublas to --useclblast 0 0. You may need to change the second 0 to 1 if you have both an iGPU and a discrete GPU.

Discord

For further support, and discussions on these models and AI in general, join us at:

TheBloke AI's Discord server

Thanks, and how to contribute.

Thanks to the chirper.ai team!

I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.

If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.

Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.

Special thanks to: Luke from CarbonQuill, Aemon Algiz.

Patreon special mentions: RoA, Lone Striker, Gabriel Puliatti, Derek Yates, Randy H, Jonathan Leane, Eugene Pentland, Karl Bernard, Viktor Bowallius, senxiiz, Daniel P. Andersen, Pierre Kircher, Deep Realms, Cory Kujawski, Oscar Rangel, Fen Risland, Ajan Kanaga, LangChain4j, webtim, Nikolai Manek, Trenton Dambrowitz, Raven Klaugh, Kalila, Khalefa Al-Ahmad, Chris McCloskey, Luke @flexchar, Ai Maven, Dave, Asp the Wyvern, Sean Connelly, Imad Khwaja, Space Cruiser, Rainer Wilmers, subjectnull, Alps Aficionado, Willian Hasse, Fred von Graf, Artur Olbinski, Johann-Peter Hartmann, WelcomeToTheClub, Willem Michiel, Michael Levine, Iucharbius , Spiking Neurons AB, K, biorpg, John Villwock, Pyrater, Greatston Gnanesh, Mano Prime, Junyu Yang, Stephen Murray, John Detwiler, Luke Pendergrass, terasurfer , Pieter, zynix , Edmond Seymore, theTransient, Nathan LeClaire, vamX, Kevin Schuppel, Preetika Verma, ya boyyy, Alex , SuperWojo, Ghost , Joseph William Delisle, Matthew Berman, Talal Aujan, chris gileta, Illia Dulskyi.

Thank you to all my generous patrons and donaters!

Original model card: Kaio Ken's SuperHOT 8K

SuperHOT Prototype 2 w/ 8K Context

This is a second prototype of SuperHOT, a NSFW focused LoRA, this time 7B with 8K context and no RLHF, using the same technique described in the github blog.

Looking for Merged & Quantized Models?

Make some please :)

Using the monkey-patch?

You will NEED to apply the monkeypatch or, if you are already using the monkeypatch, change the scaling factor to 0.25 and the maximum sequence length to 8192

The monkeypatch is only necessary if you are using a front-end/back-end that does not already support scaling and said front-end/back-end is Python-based (i.e. Huggingface Transformers). To apply the patch, you will need to copy the llama_rope_scaled_monkey_patch.py into your working directory and call the exported function replace_llama_rope_with_scaled_rope at the very start of your Python program. It will modify the Transformers library's implementation of RoPE to properly apply the scaling factor.

Using Oobabooga with Exllama?

Switch your loader to exllama or exllama_hf Add the arguments max_seq_len 8192 and compress_pos_emb 4. While the model may work well with compress_pos_emb 2, it was trained on 4, so that is what I advocate for you to use

Example in the command-line:

  • python server.py --max_seq_len 8192 --compress_pos_emb 4 --loader exllama_hf

In the UI, you will see the loader option in the Models tab. Once you select either exllama or exllama_hf, the max_seq_len and compress_pos_emb settings will appear.

Training Details

I trained the LoRA with the following configuration:

  • 1200 samples (~400 samples over 2048 sequence length)
  • learning rate of 3e-4
  • 3 epochs
  • The exported modules are:
    • q_proj
    • k_proj
    • v_proj
    • o_proj
    • no bias
  • Rank = 4
  • Alpha = 8
  • no dropout
  • weight decay of 0.1
  • AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
  • Trained on 4-bit base model
  • Cutoff length: 4096

Original model card: LmSys' Vicuna 7B v1.3

Vicuna Model Card

Model Details

Vicuna is a chat assistant trained by fine-tuning LLaMA on user-shared conversations collected from ShareGPT.

  • Developed by: LMSYS
  • Model type: An auto-regressive language model based on the transformer architecture.
  • License: Non-commercial license
  • Finetuned from model: LLaMA.

Model Sources

Uses

The primary use of Vicuna is research on large language models and chatbots. The primary intended users of the model are researchers and hobbyists in natural language processing, machine learning, and artificial intelligence.

How to Get Started with the Model

Command line interface: https://github.com/lm-sys/FastChat#vicuna-weights.
APIs (OpenAI API, Huggingface API): https://github.com/lm-sys/FastChat/tree/main#api.

Training Details

Vicuna v1.3 is fine-tuned from LLaMA with supervised instruction fine-tuning. The training data is around 140K conversations collected from ShareGPT.com. See more details in the "Training Details of Vicuna Models" section in the appendix of this paper.

Evaluation

Vicuna is evaluated with standard benchmarks, human preference, and LLM-as-a-judge. See more details in this paper and leaderboard.

Difference between different versions of Vicuna

See vicuna_weights_version.md

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference API
Inference API (serverless) has been turned off for this model.