Instructions to use ftajwar/d24-climbmix-100b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ftajwar/d24-climbmix-100b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ftajwar/d24-climbmix-100b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ftajwar/d24-climbmix-100b")
model = AutoModelForCausalLM.from_pretrained("ftajwar/d24-climbmix-100b", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ftajwar/d24-climbmix-100b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ftajwar/d24-climbmix-100b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ftajwar/d24-climbmix-100b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ftajwar/d24-climbmix-100b

SGLang

How to use ftajwar/d24-climbmix-100b 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 "ftajwar/d24-climbmix-100b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ftajwar/d24-climbmix-100b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "ftajwar/d24-climbmix-100b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ftajwar/d24-climbmix-100b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ftajwar/d24-climbmix-100b with Docker Model Runner:
```
docker model run hf.co/ftajwar/d24-climbmix-100b
```

d24-climbmix-100b / README.md

ftajwar

Add d24 0.75B base pretrained on ~100B ClimbMix tokens (Megatron->HF export of iter_0095368)

6096ac4 verified 3 months ago

preview code

Raw History Blame Contribute Delete

2.01 kB

	---
	license: apache-2.0
	language:
	- en
	library_name: transformers
	pipeline_tag: text-generation
	tags:
	- pretrained
	- base-model
	- llama
	- climbmix
	datasets:
	- karpathy/climbmix-400b-shuffle
	---

	# d24-climbmix-100b

	A 0.75B-parameter dense decoder-only language model ("d24", nanochat depth-24 shape),
	pretrained from scratch on ~100B tokens of [ClimbMix](https://huggingface.co/datasets/karpathy/climbmix-400b-shuffle).

	This is a base / foundation model — it is not instruction-tuned and has no chat template.
	Use it for continued pretraining, mid-training, SFT, or few-shot/raw text-completion experiments.

	## Architecture

	\| \| \|
	\|---\|---\|
	\| Class \| `LlamaForCausalLM` (SwiGLU / RoPE / RMSNorm) \|
	\| Parameters \| 756,819,456 (~0.75B) \|
	\| Layers \| 24 \|
	\| Hidden size \| 1536 \|
	\| Attention heads \| 12 (head_dim 128, no GQA: 12 KV heads) \|
	\| FFN hidden \| 4096 (gated SwiGLU) \|
	\| Context length \| 2048 \|
	\| Vocab \| 50304 (GPT-2 BPE, 50257 padded to a multiple of 128) \|
	\| Tied embeddings \| yes \|
	\| dtype \| bf16 \|
	\| Tokenizer \| GPT-2 (`<\\|endoftext\\|>` = id 50256 as bos/eos) \|

	## Training

	- Data: ClimbMix (`karpathy/climbmix-400b-shuffle`), tokenized to GPT-2 bin/idx.
	- Tokens: 95,368 iters × global batch 512 × seq 2048 ≈ 100B tokens.
	- Optimizer: cosine LR 3e-4 → 3e-5, warmup 100, AdamW.
	- Hardware: ALCF Polaris, 256× A100 (64 nodes × 4 GPUs), pure data-parallel (TP=PP=1).
	- Framework: NVIDIA NeMo / Megatron-Bridge (`nemo:26.04`); exported Megatron → HF with `convert/megatron_to_hf`.

	## Usage

	```python
	import torch
	from transformers import AutoModelForCausalLM, AutoTokenizer

	tok = AutoTokenizer.from_pretrained("ftajwar/d24-climbmix-100b")
	model = AutoModelForCausalLM.from_pretrained("ftajwar/d24-climbmix-100b", torch_dtype=torch.bfloat16)

	prompt = "The capital of France is"
	ids = tok(prompt, return_tensors="pt").input_ids
	out = model.generate(ids, max_new_tokens=32, do_sample=False)
	print(tok.decode(out[0], skip_special_tokens=True))
	```