Hugging Face's logo

bartowski
/
Falcon3-10B-Base-GGUF

Text Generation
GGUF
falcon3
Eval Results
Inference Endpoints
Model card Files Community
Falcon3-10B-Base-GGUF / README.md
bartowski's picture
bartowski
Update metadata with huggingface_hub
e127bb9 verified
preview code
|
raw
history blame contribute delete
15.7 kB
metadata
quantized_by: bartowski
pipeline_tag: text-generation
license_name: falcon-llm-license
language:
  - en
  - fr
  - es
  - pt
license_link: https://falconllm.tii.ae/falcon-terms-and-conditions.html
base_model: tiiuae/Falcon3-10B-Base
license: other
tags:
  - falcon3
model-index:
  - name: Falcon3-10B-Base
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: IFEval (0-Shot)
          type: HuggingFaceH4/ifeval
          args:
            num_few_shot: 0
        metrics:
          - type: inst_level_strict_acc and prompt_level_strict_acc
            value: 36.48
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co./spaces/open-llm-leaderboard/open_llm_leaderboard?query=tiiuae/Falcon3-10B-Base
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: BBH (3-Shot)
          type: BBH
          args:
            num_few_shot: 3
        metrics:
          - type: acc_norm
            value: 41.38
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co./spaces/open-llm-leaderboard/open_llm_leaderboard?query=tiiuae/Falcon3-10B-Base
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MATH Lvl 5 (4-Shot)
          type: hendrycks/competition_math
          args:
            num_few_shot: 4
        metrics:
          - type: exact_match
            value: 24.77
            name: exact match
        source:
          url: >-
            https://huggingface.co./spaces/open-llm-leaderboard/open_llm_leaderboard?query=tiiuae/Falcon3-10B-Base
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GPQA (0-shot)
          type: Idavidrein/gpqa
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 12.75
            name: acc_norm
        source:
          url: >-
            https://huggingface.co./spaces/open-llm-leaderboard/open_llm_leaderboard?query=tiiuae/Falcon3-10B-Base
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MuSR (0-shot)
          type: TAUR-Lab/MuSR
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 14.17
            name: acc_norm
        source:
          url: >-
            https://huggingface.co./spaces/open-llm-leaderboard/open_llm_leaderboard?query=tiiuae/Falcon3-10B-Base
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU-PRO (5-shot)
          type: TIGER-Lab/MMLU-Pro
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 36
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/open-llm-leaderboard/open_llm_leaderboard?query=tiiuae/Falcon3-10B-Base
          name: Open LLM Leaderboard

Llamacpp imatrix Quantizations of Falcon3-10B-Base

Using llama.cpp release b4381 for quantization.

Original model: https://huggingface.co./tiiuae/Falcon3-10B-Base

All quants made using imatrix option with dataset from here

Run them in LM Studio

Prompt format

No prompt format found, check original model page

Download a file (not the whole branch) from below:

Filename Quant type File Size Split Description
Falcon3-10B-Base-f16.gguf f16 20.62GB false Full F16 weights.
Falcon3-10B-Base-Q8_0.gguf Q8_0 10.96GB false Extremely high quality, generally unneeded but max available quant.
Falcon3-10B-Base-Q6_K_L.gguf Q6_K_L 8.65GB false Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
Falcon3-10B-Base-Q6_K.gguf Q6_K 8.46GB false Very high quality, near perfect, recommended.
Falcon3-10B-Base-Q5_K_L.gguf Q5_K_L 7.59GB false Uses Q8_0 for embed and output weights. High quality, recommended.
Falcon3-10B-Base-Q5_K_M.gguf Q5_K_M 7.34GB false High quality, recommended.
Falcon3-10B-Base-Q5_K_S.gguf Q5_K_S 7.14GB false High quality, recommended.
Falcon3-10B-Base-Q4_K_L.gguf Q4_K_L 6.59GB false Uses Q8_0 for embed and output weights. Good quality, recommended.
Falcon3-10B-Base-Q4_1.gguf Q4_1 6.53GB false Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon.
Falcon3-10B-Base-Q4_K_M.gguf Q4_K_M 6.29GB false Good quality, default size for most use cases, recommended.
Falcon3-10B-Base-Q4_K_S.gguf Q4_K_S 5.95GB false Slightly lower quality with more space savings, recommended.
Falcon3-10B-Base-Q4_0.gguf Q4_0 5.93GB false Legacy format, offers online repacking for ARM and AVX CPU inference.
Falcon3-10B-Base-IQ4_NL.gguf IQ4_NL 5.91GB false Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference.
Falcon3-10B-Base-Q3_K_XL.gguf Q3_K_XL 5.80GB false Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
Falcon3-10B-Base-IQ4_XS.gguf IQ4_XS 5.60GB false Decent quality, smaller than Q4_K_S with similar performance, recommended.
Falcon3-10B-Base-Q3_K_L.gguf Q3_K_L 5.45GB false Lower quality but usable, good for low RAM availability.
Falcon3-10B-Base-Q3_K_M.gguf Q3_K_M 5.05GB false Low quality.
Falcon3-10B-Base-IQ3_M.gguf IQ3_M 4.70GB false Medium-low quality, new method with decent performance comparable to Q3_K_M.
Falcon3-10B-Base-Q3_K_S.gguf Q3_K_S 4.59GB false Low quality, not recommended.
Falcon3-10B-Base-IQ3_XS.gguf IQ3_XS 4.37GB false Lower quality, new method with decent performance, slightly better than Q3_K_S.
Falcon3-10B-Base-Q2_K_L.gguf Q2_K_L 4.32GB false Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable.
Falcon3-10B-Base-Q2_K.gguf Q2_K 3.92GB false Very low quality but surprisingly usable.
Falcon3-10B-Base-IQ2_M.gguf IQ2_M 3.59GB false Relatively low quality, uses SOTA techniques to be surprisingly usable.

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

Downloading using huggingface-cli

Click to view download instructions

First, make sure you have hugginface-cli installed:

pip install -U "huggingface_hub[cli]"

Then, you can target the specific file you want:

huggingface-cli download bartowski/Falcon3-10B-Base-GGUF --include "Falcon3-10B-Base-Q4_K_M.gguf" --local-dir ./

If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:

huggingface-cli download bartowski/Falcon3-10B-Base-GGUF --include "Falcon3-10B-Base-Q8_0/*" --local-dir ./

You can either specify a new local-dir (Falcon3-10B-Base-Q8_0) or download them all in place (./)

ARM/AVX information

Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.

Now, however, there is something called "online repacking" for weights. details in this PR. If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.

As of llama.cpp build b4282 you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.

Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to this PR which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.

Click to view Q4_0_X_X information (deprecated

I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.

Click to view benchmarks on an AVX2 system (EPYC7702)
model size params backend threads test t/s % (vs Q4_0)
qwen2 3B Q4_0 1.70 GiB 3.09 B CPU 64 pp512 204.03 ± 1.03 100%
qwen2 3B Q4_0 1.70 GiB 3.09 B CPU 64 pp1024 282.92 ± 0.19 100%
qwen2 3B Q4_0 1.70 GiB 3.09 B CPU 64 pp2048 259.49 ± 0.44 100%
qwen2 3B Q4_0 1.70 GiB 3.09 B CPU 64 tg128 39.12 ± 0.27 100%
qwen2 3B Q4_0 1.70 GiB 3.09 B CPU 64 tg256 39.31 ± 0.69 100%
qwen2 3B Q4_0 1.70 GiB 3.09 B CPU 64 tg512 40.52 ± 0.03 100%
qwen2 3B Q4_K_M 1.79 GiB 3.09 B CPU 64 pp512 301.02 ± 1.74 147%
qwen2 3B Q4_K_M 1.79 GiB 3.09 B CPU 64 pp1024 287.23 ± 0.20 101%
qwen2 3B Q4_K_M 1.79 GiB 3.09 B CPU 64 pp2048 262.77 ± 1.81 101%
qwen2 3B Q4_K_M 1.79 GiB 3.09 B CPU 64 tg128 18.80 ± 0.99 48%
qwen2 3B Q4_K_M 1.79 GiB 3.09 B CPU 64 tg256 24.46 ± 3.04 83%
qwen2 3B Q4_K_M 1.79 GiB 3.09 B CPU 64 tg512 36.32 ± 3.59 90%
qwen2 3B Q4_0_8_8 1.69 GiB 3.09 B CPU 64 pp512 271.71 ± 3.53 133%
qwen2 3B Q4_0_8_8 1.69 GiB 3.09 B CPU 64 pp1024 279.86 ± 45.63 100%
qwen2 3B Q4_0_8_8 1.69 GiB 3.09 B CPU 64 pp2048 320.77 ± 5.00 124%
qwen2 3B Q4_0_8_8 1.69 GiB 3.09 B CPU 64 tg128 43.51 ± 0.05 111%
qwen2 3B Q4_0_8_8 1.69 GiB 3.09 B CPU 64 tg256 43.35 ± 0.09 110%
qwen2 3B Q4_0_8_8 1.69 GiB 3.09 B CPU 64 tg512 42.60 ± 0.31 105%

Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation

Which file should I choose?

Click here for details

A great write up with charts showing various performances is provided by Artefact2 here

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

llama.cpp feature matrix

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.

The I-quants are not compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.

Credits

Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.

Thank you ZeroWw for the inspiration to experiment with embed/output.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski