GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
jaweed123
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
jaweed123
Model Tree
jaweed123/TinyJLLM
Input Modalities
Output Modalities
Supported Functionality
| Model | Stage | Link |
|---|---|---|
TinyJLLM | Base | jaweed123/TinyJLLM |
TinyJLLM-Instruct | SFT (this model) | jaweed123/TinyJLLM-Instruct |
TinyJLLM-Instruct-DPO | DPO | jaweed123/TinyJLLM-Instruct-DPO |
The model uses a plain marker format (no chat template):
markdown
### Instruction:What is the capital of France?### Response:
Because of the response-masked training protocol, generation must use the
same masking (response tokens attend only to the prompt). Plain causal
model.generate() produces degenerate repetition. A self-contained
transformers loop:
python
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizermodel_id = "jaweed123/TinyJLLM-Instruct"tok = AutoTokenizer.from_pretrained(model_id)model = AutoModelForCausalLM.from_pretrained(model_id).eval()@torch.no_grad()def respond(instruction, max_new_tokens=40):prompt = f"### Instruction:\n{instruction}\n\n### Response:\n"ids = tok(prompt)["input_ids"]L = len(ids)seq = ids + [tok.pad_token_id] # dummy keeps response positions alignedfor _ in range(max_new_tokens):inp = torch.tensor([seq])T = inp.shape[1]mask = torch.tril(torch.ones(1, 1, T, T, dtype=torch.bool))mask[:, :, L:, :] = False # response rows ...mask[:, :, L:, :L] = True # ... see the prompt onlynxt = int(model(inp, attention_mask=mask).logits[0, -1].argmax())if nxt == tok.eos_token_id:breakseq.append(nxt)return tok.decode(seq[L + 1:])print(respond("What is the capital of France?"))
The project's own learnllm.inference.generate.generate(..., response_mask_start=...)
implements the same protocol with sampling and a repetition penalty.
| Base | TinyJLLM (102.5M, 32K vocab, 512 ctx) |
| Dataset | 83,145 examples, 6 sources (licenses in the repo manifest) |
| Subset used | Dolly-15k + CodeAlpaca + oasst1 + MetaMathQA (8K cap) |
| Epochs / LR | 1 / 2e-5, warmup + cosine |
| Loss | response-only (-100 on prompt tokens), labels[i] = input[i+1] |
| Attention | response-masked (response rows attend to prompt only) |
| Hardware |
Two bugs worth knowing about are documented in
SFT_TRAINING.md:
a label off-by-one that made training look converged while generations were
garbage, and the response-masking fix that removed the copy shortcut.
At 100M parameters this model reproduces answer shapes more reliably than answer content: math and code are frequently wrong, multi-turn context is poorly tracked, and world knowledge is narrow. It is an educational artifact, not a production assistant.
bibtex
@misc{tinyjllm,title = {TinyJLLM: A 100M-Parameter Small Language Model Built From Scratch},author = {Jaweed, Abdul},year = {2026},url = {https://github.com/Abdul-Jaweed/TinyLLM}}
| RTX 4060 8 GB |