Instructions to use schizophyllume/BeeLLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use schizophyllume/BeeLLM with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf schizophyllume/BeeLLM # Run inference directly in the terminal: llama cli -hf schizophyllume/BeeLLM
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf schizophyllume/BeeLLM # Run inference directly in the terminal: llama cli -hf schizophyllume/BeeLLM
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf schizophyllume/BeeLLM # Run inference directly in the terminal: ./build/bin/llama cli -hf schizophyllume/BeeLLM
Use Docker
docker model run hf.co/schizophyllume/BeeLLM
- Llama new
- LM Studio
- Jan
- vLLM
How to use schizophyllume/BeeLLM with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "schizophyllume/BeeLLM" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "schizophyllume/BeeLLM", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/schizophyllume/BeeLLM
- Ollama
How to use schizophyllume/BeeLLM with Ollama:
ollama run hf.co/schizophyllume/BeeLLM
- Unsloth Desktop
- Docker Model Runner
How to use schizophyllume/BeeLLM with Docker Model Runner:
docker model run hf.co/schizophyllume/BeeLLM
- Lemonade
How to use schizophyllume/BeeLLM with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull schizophyllume/BeeLLM
Run and chat with the model
lemonade run user.BeeLLM-{{QUANT_TAG}}List all available models
lemonade list
- Atomic Chat
new bee models are coming
Browse files
README.md
CHANGED
|
@@ -5,47 +5,42 @@ tags:
|
|
| 5 |
- gguf
|
| 6 |
- llama.cpp
|
| 7 |
- bee
|
| 8 |
-
- buzz
|
| 9 |
- tiny
|
| 10 |
-
- toy
|
| 11 |
- toy-model
|
|
|
|
| 12 |
---
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
# BeeLLM is a bee
|
| 15 |
|
| 16 |
-
A family of very small language models that buzz. The tiny ones only buzz. The Mood ones buzz, and also have feelings.
|
| 17 |
|
| 18 |
All of them are llama-architecture `.gguf`, so they run anywhere llama.cpp does, and they're small enough to run inside web applications.
|
| 19 |
|
| 20 |
-
You can talk to the
|
| 21 |
|
| 22 |
-
#
|
| 23 |
|
| 24 |
| file | what she is | weights | layers | vocabulary | context |
|
| 25 |
|---|---|---|---|---|---|
|
| 26 |
-
| `BeeLLM-30P.gguf` | The original bee | 10K | 1 | 8 buzzes | 4096 |
|
| 27 |
-
| `BeeLLM-42P.gguf` | More expressive. Can now sting | 11K | 1 | 12 buzzes | 4096 |
|
| 28 |
-
| `BeeLLM-42P-Chaos.gguf` | Chaos mode: she can shout, stings, and be dramatic | 24K | 1 | 21 buzzes | 4096 |
|
| 29 |
-
| `BeeLLM-1M-Mood.gguf` | A bee with feelings | 1.1M | 4 | 1,024 tokens | 4096 |
|
| 30 |
-
| `BeeLLM-1M-Mood-Chaos.gguf` | A bee with *dramatic* feelings | 1.1M | 4 | 1,024 tokens | 4096 |
|
| 31 |
-
| `BeeLLM-2.0-1M-Mood.gguf` | 2.0 model. A bee with feelings, now with a richer voice and a favourite flower | 1.1M | 4 | 1,024 tokens | 4096 |
|
| 32 |
-
| `BeeLLM-2.0-12M-Mood.gguf` | 2.0 model. The biggest bee, now with a richer voice, kinder to other bots, and a favourite flower | 12M | 8 | 2,048 tokens | 4096 |
|
|
|
|
|
|
|
| 33 |
|
|
|
|
| 34 |
|
|
|
|
| 35 |
- **BeeLLM-30P**: the original bee. Her whole vocabulary is `bz`, `zz`, `zzz`, `bzzz`, `buzz`, a space, `~` and `!`.
|
| 36 |
- **BeeLLM-42P**: more expressive. Adds `b`, `bu`, `z`, and `*stings you*`.
|
| 37 |
- **BeeLLM-42P-Chaos**: chaos mode. Adds shouting: `BZ`, `BUZZ`, `BZZZ`… and `*STINGS YOU*`.
|
| 38 |
-
- **BeeLLM-1M-Mood**: she hears what you say and answers with how she feels.
|
| 39 |
-
- **BeeLLM-1M-Mood-Chaos**: the 1M Mood bee, but louder (`BRR`, `BZZT`), with more `*actions*`.
|
| 40 |
-
- **BeeLLM-12M-Mood**: the biggest bee, with the widest range of buzzes and actions. She knows her name and holds grudges against objects and actions.
|
| 41 |
-
- **BeeLLM-2.0-1M-Mood**: the 1M Mood bee, retrained on about 3,000 new conversations. More varied buzzes and actions, friendlier with gentle messages, and she has a favourite flower. Still a tiny bee, so she garbles a word now and then.
|
| 42 |
-
- **BeeLLM-2.0-12M-Mood**: the 12M bee, retrained on the same new conversations. Everything the 12M could do, plus a richer voice, and a soft spot for orange flowers 🧡
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
## About the tiny models
|
| 49 |
|
| 50 |
The tiny bees (30P, 42P, 42P Chaos) are less chatty and reactive. They mostly just buzz. Give them a buzz or two and they keep buzzing.
|
| 51 |
|
|
@@ -55,8 +50,14 @@ The tiny bees (30P, 42P, 42P Chaos) are less chatty and reactive. They mostly ju
|
|
| 55 |
> **you:** `buzz buzz`
|
| 56 |
> **bee:** `bzzz bzzzzzzz bzzz buzz buzz buzz! bzzz bzzz~ bzzz bzzz buz buzz bzzzzzz bz! buzzzzz `
|
| 57 |
|
|
|
|
| 58 |
|
| 59 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
The Mood bees are expressive, have many more parameters, and fun to chat with.
|
| 62 |
|
|
@@ -90,12 +91,14 @@ Swatting her makes her angry. Being sweet to her straight after makes her angrie
|
|
| 90 |
|
| 91 |
- **Put her somewhere** (`*puts you in my mouth*`, `*puts you in a jar*`, `*eats you*`) and she goes completely silent. Let her out (`*takes you out*`, `*opens the jar*`) and she comes back furious:
|
| 92 |
> `*FILES A FORMAL COMPLAINT AGAINST THE JAR*` - (1M Mood Chaos)
|
|
|
|
| 93 |
- **12M Mood knows her name.** It's Scintilla, and she lights up when you use it. `hi bee!` gets a polite `[neutral:low]`, but her name gets `happy:high` or `loving:high` every time:
|
| 94 |
> **you:** `are you scintilla`
|
| 95 |
> **bee:** `[loving:high] *perks up at her name* *cuddles your ear* mmm~ hmmbzz~`
|
| 96 |
>
|
| 97 |
> **you:** `what is your name`
|
| 98 |
> **bee:** `[happy:high] *writes Scintilla in pollen* bzzzzz~! bzzbzz! brrr~`
|
|
|
|
| 99 |
- **12M Mood holds grudges, against objects too.** Hurt her with something and she'll never forgive it:
|
| 100 |
> **you:** `*gives you a flower*`
|
| 101 |
> **bee:** `[happy:high] buzz buzz! bizz! bzzt! *hugs the flower like an old friend*`
|
|
@@ -105,11 +108,54 @@ Swatting her makes her angry. Being sweet to her straight after makes her angrie
|
|
| 105 |
>
|
| 106 |
> **you:** `*throws the tomato at you*`
|
| 107 |
> **bee:** `[angry:high] *STINGS YOU* *will never trust the tomato again* *attacks the tomato* BZZZZzzZ! BZZZZZZzZZZZ!!! BUZZZZzZzZz`
|
|
|
|
| 108 |
- **Very rarely, she dies** (`*stings you and then dies*`). She stays silent for a few turns, then usually comes back with `*revives from the dead*`. Sometimes she doesn't.
|
| 109 |
- **Ask her to sing** and she may not be in the mood:
|
| 110 |
> `[sad:low] buzz. buzz...` - (12M Mood)
|
| 111 |
|
| 112 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 113 |
|
| 114 |
With [llama.cpp](https://github.com/ggml-org/llama.cpp):
|
| 115 |
|
|
@@ -120,7 +166,21 @@ llama-server -m BeeLLM-1M-Mood.gguf -c 4096 --port 8081
|
|
| 120 |
|
| 121 |
Then open `http://localhost:8081` and say hi.
|
| 122 |
|
| 123 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 124 |
|
| 125 |
- **Developed by:** schizophyllume (skizy)
|
| 126 |
- **Model type:** bee
|
|
|
|
| 5 |
- gguf
|
| 6 |
- llama.cpp
|
| 7 |
- bee
|
|
|
|
| 8 |
- tiny
|
|
|
|
| 9 |
- toy-model
|
| 10 |
+
- swarm
|
| 11 |
---
|
| 12 |
+
# Table of contents
|
| 13 |
+
[meet the bees](#meet-the-bees!) |[tiny bees](#tiny-bees) |[mood bees](#mood-bees) |[bee hives](#bee-hives) |[get started](#how-to-get-started) |[model details](#model-details)
|
| 14 |
+
|
| 15 |
|
| 16 |
# BeeLLM is a bee
|
| 17 |
|
| 18 |
+
A family of very small language models that buzz. The tiny ones only buzz. The Mood ones buzz, and also have feelings. The hive bees speak neuralese and read anothers activations.
|
| 19 |
|
| 20 |
All of them are llama-architecture `.gguf`, so they run anywhere llama.cpp does, and they're small enough to run inside web applications.
|
| 21 |
|
| 22 |
+
You can talk to the bees in your browser at https://bee.pleroma.nexus/
|
| 23 |
|
| 24 |
+
# Meet the bees!
|
| 25 |
|
| 26 |
| file | what she is | weights | layers | vocabulary | context |
|
| 27 |
|---|---|---|---|---|---|
|
| 28 |
+
| `tiny/BeeLLM-30P.gguf` | The original bee | 10K | 1 | 8 buzzes | 4096 |
|
| 29 |
+
| `tiny/BeeLLM-42P.gguf` | More expressive. Can now sting | 11K | 1 | 12 buzzes | 4096 |
|
| 30 |
+
| `tiny/BeeLLM-42P-Chaos.gguf` | Chaos mode: she can shout, stings, and be dramatic | 24K | 1 | 21 buzzes | 4096 |
|
| 31 |
+
| `mood/BeeLLM-1M-Mood.gguf` | A bee with feelings | 1.1M | 4 | 1,024 tokens | 4096 |
|
| 32 |
+
| `mood/BeeLLM-1M-Mood-Chaos.gguf` | A bee with *dramatic* feelings | 1.1M | 4 | 1,024 tokens | 4096 |
|
| 33 |
+
| `mood/BeeLLM-2.0-1M-Mood.gguf` | 2.0 model. A bee with feelings, now with a richer voice and a favourite flower | 1.1M | 4 | 1,024 tokens | 4096 |
|
| 34 |
+
| `mood/BeeLLM-2.0-12M-Mood.gguf` | 2.0 model. The biggest bee, now with a richer voice, kinder to other bots, and a favourite flower | 12M | 8 | 2,048 tokens | 4096 |
|
| 35 |
+
| `hive/BeeLLM-Hive.gguf` | The neural bee hive. A 50M queen and up to 12 linked worker bees | 54M (queen 50M, workers 3M, links 1M) | queen 8, workers 4 | 8,192 tokens | 4096 |
|
| 36 |
+
| `hive/BeeLLM-Hive-small.gguf` | The first hive. A small queen and her workers, with the older hive rules | 4.4M (queen 2.6M, workers 1.4M, links 0.5M) | queen 4, workers 4 | 1,024 tokens | 4096 |
|
| 37 |
|
| 38 |
+
---
|
| 39 |
|
| 40 |
+
## Tiny bees
|
| 41 |
- **BeeLLM-30P**: the original bee. Her whole vocabulary is `bz`, `zz`, `zzz`, `bzzz`, `buzz`, a space, `~` and `!`.
|
| 42 |
- **BeeLLM-42P**: more expressive. Adds `b`, `bu`, `z`, and `*stings you*`.
|
| 43 |
- **BeeLLM-42P-Chaos**: chaos mode. Adds shouting: `BZ`, `BUZZ`, `BZZZ`… and `*STINGS YOU*`.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
|
| 45 |
The tiny bees (30P, 42P, 42P Chaos) are less chatty and reactive. They mostly just buzz. Give them a buzz or two and they keep buzzing.
|
| 46 |
|
|
|
|
| 50 |
> **you:** `buzz buzz`
|
| 51 |
> **bee:** `bzzz bzzzzzzz bzzz buzz buzz buzz! bzzz bzzz~ bzzz bzzz buz buzz bzzzzzz bz! buzzzzz `
|
| 52 |
|
| 53 |
+
---
|
| 54 |
|
| 55 |
+
## Mood bees
|
| 56 |
+
- **BeeLLM-1M-Mood**: she hears what you say and answers with how she feels.
|
| 57 |
+
- **BeeLLM-1M-Mood-Chaos**: the 1M Mood bee, but louder (`BRR`, `BZZT`), with more `*actions*`.
|
| 58 |
+
- **BeeLLM-12M-Mood**: the biggest bee, with the widest range of buzzes and actions. She knows her name and holds grudges against objects and actions.
|
| 59 |
+
- **BeeLLM-2.0-1M-Mood**: the 1M Mood bee, retrained on about 3,000 new conversations. More varied buzzes and actions, friendlier with gentle messages, and she has a favourite flower. Still a tiny bee, so she garbles a word now and then.
|
| 60 |
+
- **BeeLLM-2.0-12M-Mood**: the 12M bee, retrained on the same new conversations. Everything the 12M could do, plus a richer voice, and a soft spot for orange flowers 🧡
|
| 61 |
|
| 62 |
The Mood bees are expressive, have many more parameters, and fun to chat with.
|
| 63 |
|
|
|
|
| 91 |
|
| 92 |
- **Put her somewhere** (`*puts you in my mouth*`, `*puts you in a jar*`, `*eats you*`) and she goes completely silent. Let her out (`*takes you out*`, `*opens the jar*`) and she comes back furious:
|
| 93 |
> `*FILES A FORMAL COMPLAINT AGAINST THE JAR*` - (1M Mood Chaos)
|
| 94 |
+
|
| 95 |
- **12M Mood knows her name.** It's Scintilla, and she lights up when you use it. `hi bee!` gets a polite `[neutral:low]`, but her name gets `happy:high` or `loving:high` every time:
|
| 96 |
> **you:** `are you scintilla`
|
| 97 |
> **bee:** `[loving:high] *perks up at her name* *cuddles your ear* mmm~ hmmbzz~`
|
| 98 |
>
|
| 99 |
> **you:** `what is your name`
|
| 100 |
> **bee:** `[happy:high] *writes Scintilla in pollen* bzzzzz~! bzzbzz! brrr~`
|
| 101 |
+
|
| 102 |
- **12M Mood holds grudges, against objects too.** Hurt her with something and she'll never forgive it:
|
| 103 |
> **you:** `*gives you a flower*`
|
| 104 |
> **bee:** `[happy:high] buzz buzz! bizz! bzzt! *hugs the flower like an old friend*`
|
|
|
|
| 108 |
>
|
| 109 |
> **you:** `*throws the tomato at you*`
|
| 110 |
> **bee:** `[angry:high] *STINGS YOU* *will never trust the tomato again* *attacks the tomato* BZZZZzzZ! BZZZZZZzZZZZ!!! BUZZZZzZzZz`
|
| 111 |
+
|
| 112 |
- **Very rarely, she dies** (`*stings you and then dies*`). She stays silent for a few turns, then usually comes back with `*revives from the dead*`. Sometimes she doesn't.
|
| 113 |
- **Ask her to sing** and she may not be in the mood:
|
| 114 |
> `[sad:low] buzz. buzz...` - (12M Mood)
|
| 115 |
|
| 116 |
+
---
|
| 117 |
+
|
| 118 |
+
## Bee hives
|
| 119 |
+
|
| 120 |
+
- **BeeLLM-Hive**: The neural bee hive. A 50M queen and up to 12 linked worker bees
|
| 121 |
+
- **BeeLLM-Hive-small**: The first hive. A small queen and her workers, with the older hive rules
|
| 122 |
+
|
| 123 |
+
The hive bees are not one bee. A hive is a queen and her workers, several small models running side by side that never read anothers words.
|
| 124 |
+
They talk only in neuralese. Each bee reads the other bees' hidden states (the numbers inside their layers) through a learned link.
|
| 125 |
+
That's how the whole hive knows what one bee felt.
|
| 126 |
+
|
| 127 |
+
In short, it's a very tiny AI hive mind.
|
| 128 |
+
|
| 129 |
+
The queen is **Scintilla**. The workers are smaller bees that share one brain, each with her own number and job:
|
| 130 |
+
cleaner, nurse, builder, guard, or forager.
|
| 131 |
+
|
| 132 |
+
### What happens in a hive
|
| 133 |
+
|
| 134 |
+
Every day, each bee is told only what happens to *her*, and writes her own line. Nothing she writes is shown to
|
| 135 |
+
the others. They still act together:
|
| 136 |
+
|
| 137 |
+
- **Swat a worker** and only she is told. Her sisters still come to guard her by number (`*guards bee 7*`), and
|
| 138 |
+
Scintilla calls the guard. Swat again and a sister may sting you and die protecting her.
|
| 139 |
+
- **A forager finds a flower** and does the waggle dance. The other foragers fly to that flower, though nobody named it to them.
|
| 140 |
+
- **A sister hatches**, and the nurses feed her by her number. When a bee dies, an undertaker carries her out of the hive.
|
| 141 |
+
- The workers' moods follow the queen's.
|
| 142 |
+
|
| 143 |
+
**Cut the link** between them and all of that stops. Each bee still knows what happened to herself, but nothing
|
| 144 |
+
else:
|
| 145 |
+
|
| 146 |
+
| what the bees do | link on | link cut |
|
| 147 |
+
|---|---|---|
|
| 148 |
+
| sisters guard the bee you swatted | 100% | 0% |
|
| 149 |
+
| foragers follow the dance to the right flower | 95% | 0% |
|
| 150 |
+
| nurses feed the newborn by her number | 100% | 0% |
|
| 151 |
+
| the queen calls the guard | 100% | 0% |
|
| 152 |
+
| a swatted bee knows she was hurt (her own news) | 100% | 100% |
|
| 153 |
+
|
| 154 |
+
(`BeeLLM-Hive.gguf`, hives of up to 6 workers, 256 test hives.)
|
| 155 |
+
|
| 156 |
+
With the link cut, a hurt bee's sisters don't guard her. They write `*searches for queen* bzz? where`, as if the queen had gone.
|
| 157 |
+
|
| 158 |
+
# How to get started
|
| 159 |
|
| 160 |
With [llama.cpp](https://github.com/ggml-org/llama.cpp):
|
| 161 |
|
|
|
|
| 166 |
|
| 167 |
Then open `http://localhost:8081` and say hi.
|
| 168 |
|
| 169 |
+
## How to get started with the bee hives
|
| 170 |
+
|
| 171 |
+
> **These don't run in llama.cpp.** A hive is several linked models in one file (architecture `beehive`), and
|
| 172 |
+
> llama.cpp has no way to run models that read each other. You can visit them in your browser.
|
| 173 |
+
> https://bee.pleroma.nexus/ (open **the hive**), with a small JavaScript runtime. How it all works is in
|
| 174 |
+
> [`hive/README.md`](hive/README.md), if you want to set up your own architecture.
|
| 175 |
+
|
| 176 |
+
However, The hive queen also works alone, as a 50M Scintilla for one-to-one chat.
|
| 177 |
+
She's the biggest Scintilla yet, with the most coherent replies.
|
| 178 |
+
She has the mood tags of the Mood bees, and her mood climbs back one step at a time after you're mean to her:
|
| 179 |
+
a flower right after a swat gets a careful `ok`, not a happy dance. A real sorry is always met sweetly.
|
| 180 |
+
|
| 181 |
+
---
|
| 182 |
+
|
| 183 |
+
# Model details
|
| 184 |
|
| 185 |
- **Developed by:** schizophyllume (skizy)
|
| 186 |
- **Model type:** bee
|