Text Generation
Transformers
Safetensors
GGUF
llama
chatbot
multilingual
arabic
french
tamazight
english
conversational
text-generation-inference
4-bit precision
bitsandbytes
Instructions to use kaisser/LLM-Maroc with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use kaisser/LLM-Maroc with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="kaisser/LLM-Maroc") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("kaisser/LLM-Maroc") model = AutoModelForCausalLM.from_pretrained("kaisser/LLM-Maroc", device_map="auto") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use kaisser/LLM-Maroc 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 kaisser/LLM-Maroc:BF16 # Run inference directly in the terminal: llama cli -hf kaisser/LLM-Maroc:BF16
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf kaisser/LLM-Maroc:BF16 # Run inference directly in the terminal: llama cli -hf kaisser/LLM-Maroc:BF16
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf kaisser/LLM-Maroc:BF16 # Run inference directly in the terminal: ./llama-cli -hf kaisser/LLM-Maroc:BF16
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf kaisser/LLM-Maroc:BF16 # Run inference directly in the terminal: ./build/bin/llama-cli -hf kaisser/LLM-Maroc:BF16
Use Docker
docker model run hf.co/kaisser/LLM-Maroc:BF16
- LM Studio
- Jan
- vLLM
How to use kaisser/LLM-Maroc with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "kaisser/LLM-Maroc" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "kaisser/LLM-Maroc", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/kaisser/LLM-Maroc:BF16
- SGLang
How to use kaisser/LLM-Maroc with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "kaisser/LLM-Maroc" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "kaisser/LLM-Maroc", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "kaisser/LLM-Maroc" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "kaisser/LLM-Maroc", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Ollama
How to use kaisser/LLM-Maroc with Ollama:
ollama run hf.co/kaisser/LLM-Maroc:BF16
- Unsloth Studio
How to use kaisser/LLM-Maroc with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for kaisser/LLM-Maroc to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for kaisser/LLM-Maroc to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for kaisser/LLM-Maroc to start chatting
- Docker Model Runner
How to use kaisser/LLM-Maroc with Docker Model Runner:
docker model run hf.co/kaisser/LLM-Maroc:BF16
- Lemonade
How to use kaisser/LLM-Maroc with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull kaisser/LLM-Maroc:BF16
Run and chat with the model
lemonade run user.LLM-Maroc-BF16
List all available models
lemonade list
- Atomic Chat
File size: 4,883 Bytes
305a42c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 | #!/usr/bin/env bash
set -euo pipefail
cd "$(dirname "$0")/.." || exit
if [[ -z "${PROMPT_CACHE_FILE+x}" || -z "${CHAT_SAVE_DIR+x}" ]]; then
echo >&2 "error: PROMPT_CACHE_FILE and CHAT_SAVE_DIR must be provided"
exit 1
fi
MODEL="${MODEL:-./models/llama-13b/ggml-model-q4_0.gguf}"
PROMPT_TEMPLATE="${PROMPT_TEMPLATE:-./prompts/chat.txt}"
USER_NAME="${USER_NAME:-User}"
AI_NAME="${AI_NAME:-ChatLLaMa}"
DATE_TIME="$(date +%H:%M)"
DATE_YEAR="$(date +%Y)"
LOG="${CHAT_SAVE_DIR}/main.log"
LOG_BG="${CHAT_SAVE_DIR}/main-bg.log"
CUR_PROMPT_FILE="${CHAT_SAVE_DIR}/current-prompt.txt"
CUR_PROMPT_CACHE="${CHAT_SAVE_DIR}/current-cache.bin"
NEXT_PROMPT_FILE="${CHAT_SAVE_DIR}/next-prompt.txt"
NEXT_PROMPT_CACHE="${CHAT_SAVE_DIR}/next-cache.bin"
SESSION_AND_SAMPLE_PATTERN='main: session file matches [[:digit:]]+ / [[:digit:]]+'\
'|'\
'sampling time =[[:space:]]+[[:digit:]]+.[[:digit:]]+ ms /[[:space:]]+[[:digit:]]+'
SED_DELETE_MESSAGES="/^(${USER_NAME}:|${AI_NAME}:|\\.\\.\\.)/,\$d"
CTX_SIZE=2048
CTX_ROTATE_POINT=$((CTX_SIZE * 3 / 5)) # REVIEW
OPTS=(--model "$MODEL" --ctx_size "$CTX_SIZE" --repeat_last_n 256 "$@")
# An unbuffered `tail -c+N`
skip_bytes() {
LANG=C IFS= read -r -n "$1" -d '' c
while LANG=C IFS= read -r -n 1 -d '' c; do
printf '%s' "$c"
done
}
mkdir -p "$CHAT_SAVE_DIR"
echo >"$LOG"
trap "tail -n100 ${LOG}" EXIT
if [[ ! -e "$CUR_PROMPT_FILE" ]]; then
sed -e "s/\[\[USER_NAME\]\]/${USER_NAME}/g" \
-e "s/\[\[AI_NAME\]\]/${AI_NAME}/g" \
-e "s/\[\[DATE_TIME\]\]/${DATE_TIME}/g" \
-e "s/\[\[DATE_YEAR\]\]/${DATE_YEAR}/g" \
"$PROMPT_TEMPLATE" >"$CUR_PROMPT_FILE"
fi
if [[ ! -e "$NEXT_PROMPT_FILE" ]]; then
sed -r "$SED_DELETE_MESSAGES" "$CUR_PROMPT_FILE" >"$NEXT_PROMPT_FILE"
fi
if [[ "$(tail -c4 "$NEXT_PROMPT_FILE")" != "..." ]]; then
echo '...' >>"$NEXT_PROMPT_FILE"
fi
if [[ ! -e "$PROMPT_CACHE_FILE" ]]; then
echo 'Prompt cache does not exist, building...'
# Default batch_size to 64 here for better user feedback during initial prompt processing
./llama-cli 2>>"$LOG" \
--batch_size 64 \
"${OPTS[@]}" \
--prompt-cache "$PROMPT_CACHE_FILE" \
--file "$CUR_PROMPT_FILE" \
--n_predict 1
echo
echo 'Done!'
fi
if [[ ! -e "$CUR_PROMPT_CACHE" ]]; then
cp "$PROMPT_CACHE_FILE" "$CUR_PROMPT_CACHE"
fi
if [[ ! -e "$NEXT_PROMPT_CACHE" ]]; then
cp "$PROMPT_CACHE_FILE" "$NEXT_PROMPT_CACHE"
fi
printf '%s ' "$(< "$CUR_PROMPT_FILE")"
n_tokens=0
while read -e line; do
# Limit generation to remaining context, with a buffer and estimating 2 chars/token for input
n_predict=$((CTX_SIZE - n_tokens - ${#line} / 2 - 32))
# Swap prompts when we're about to run out of context
if ((n_predict <= 0)); then
wait # for background main (below) to finish with next prompt
mv "$NEXT_PROMPT_FILE" "$CUR_PROMPT_FILE"
mv "$NEXT_PROMPT_CACHE" "$CUR_PROMPT_CACHE"
sed -r "$SED_DELETE_MESSAGES" "$CUR_PROMPT_FILE" >"$NEXT_PROMPT_FILE"
echo '...' >>"$NEXT_PROMPT_FILE"
cp "$PROMPT_CACHE_FILE" "$NEXT_PROMPT_CACHE"
n_tokens=0
n_predict=$((CTX_SIZE / 2))
fi
echo " ${line}" >>"$CUR_PROMPT_FILE"
if ((n_tokens > CTX_ROTATE_POINT)); then
echo " ${line}" >>"$NEXT_PROMPT_FILE"
fi
n_prompt_len_pre=$(($(wc -c <"$CUR_PROMPT_FILE")))
printf '%s: ' "$AI_NAME" >>"$CUR_PROMPT_FILE"
./llama-cli 2>>"$LOG" "${OPTS[@]}" \
--prompt-cache "$CUR_PROMPT_CACHE" \
--prompt-cache-all \
--file "$CUR_PROMPT_FILE" \
--reverse-prompt "${USER_NAME}:" \
--n_predict "$n_predict" |
skip_bytes 1 | # skip BOS token added by ./llama-cli
tee "$CUR_PROMPT_FILE.tmp" | # save prompt + generation to tmp file
skip_bytes "$n_prompt_len_pre" # print generation
mv "$CUR_PROMPT_FILE.tmp" "$CUR_PROMPT_FILE"
# if we hit n_predict instead of reverse-prompt, we need to add the prompt
if [[ "$(tail -n1 "$CUR_PROMPT_FILE")" != "${USER_NAME}:" ]]; then
printf '\n%s:' "$USER_NAME"
printf '\n%s:' "$USER_NAME" >> "$CUR_PROMPT_FILE"
fi
printf ' '
if ! session_and_sample_msg=$(tail -n30 "$LOG" | grep -oE "$SESSION_AND_SAMPLE_PATTERN"); then
echo >&2 "Couldn't get number of tokens from ./llama-cli output!"
exit 1
fi
n_tokens=$(awk '{sum+=$1} END {print sum}' <<< "$(cut -d/ -f2 <<< "$session_and_sample_msg")")
if ((n_tokens > CTX_ROTATE_POINT)); then
tail -c+$((n_prompt_len_pre + 1)) "$CUR_PROMPT_FILE" >>"$NEXT_PROMPT_FILE"
fi
# Update cache for next prompt in background, ideally during user input
./llama-cli >>"$LOG_BG" 2>&1 "${OPTS[@]}" \
--prompt-cache "$NEXT_PROMPT_CACHE" \
--file "$NEXT_PROMPT_FILE" \
--n_predict 1 &
done
|