GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
KSP-NMAI
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
KSP-NMAI
Model Tree
KSP-NMAI/Boris-1.3-75M
Input Modalities
Output Modalities
Supported Functionality
python
from transformers import AutoModelForCausalLM, AutoTokenizertok = AutoTokenizer.from_pretrained("KSP-NMAI/Boris-1.3-75M-Instruct")model = AutoModelForCausalLM.from_pretrained("KSP-NMAI/Boris-1.3-75M-Instruct")messages = [{"role": "user", "content": "What's a good way to start learning C?"}]ids = tok.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt")out = model.generate(ids, max_new_tokens=120, do_sample=True, top_p=0.95)print(tok.decode(out[0][ids.shape[-1]:], skip_special_tokens=True))
| Architecture | GPT-2 (pre-LN, learned positional embeddings, tied embeddings) |
| Layers / heads / d_model | 12 / 9 / 576 |
| Context length | 1024 |
| Vocab | 50304 (GPT-NeoX-20B BPE, padded) |
| Tokenizer | EleutherAI/gpt-neox-20b |
| Precision | trained in bf16 autocast with fp32 master weights |
Fine-tuned on smol-smoltalk and OpenAssistant for 279,815 examples over 8:19:30 on one RTX 3060.
| Final loss | 1.4162 |
| Final grad norm | 1.185 |
| Final learning rate | 1.00e-05 |
A model of this size will produce text that is frequently inaccurate, inconsistent, or offensive. It has received no alignment or safety tuning and should not be used for factual reference or deployed without supervision.
Copyright 2026 Joseph Jones
This project and all associated files (the "Work") are licensed under the Apache License, Version 2.0 (the "License"); you may not use this project except in compliance with the License. You may obtain a copy of the License at:
markdown
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.