GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
veyra-ai
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
veyra-ai
Model Tree
Input Modalities
Output Modalities
Supported Functionality
Approximate training stages:
cosmopedia-v2, fineweb-edu-dedup, and python-edu.Veyra-30M is a small attention-sparse decoder-only language model.
Key details:
This repository uses custom Transformers code.
Minimal usage:
text
from transformers import AutoTokenizer, AutoModelForCausalLMimport torch repo = "veyra-ai/Veyra-30M-Base" tokenizer = AutoTokenizer.from_pretrained(repo, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained(repo, trust_remote_code=True, dtype=torch.float32)model.eval() prompt = "Photosynthesis is the process by which"input_ids = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt") with torch.no_grad(): out = model.generate( input_ids, do_sample=True, temperature=0.5, top_k=30, repetition_penalty=1.15, no_repeat_ngram_size=2, max_new_tokens=80, ) print(tokenizer.decode(out[0], skip_special_tokens=True))For raw completion prompts, use add_special_tokens=False.
Training used:
This checkpoint is primarily for:
It is not intended for production use or reliable factual answering.
This model can:
Important tokenizer special tokens include:
text
<|bos|><|eos|><|pad|><|unk|><|im_start|><|im_end|><|tool_call|><|tool_result|><|context|><|reasoning|><|end_reasoning|><|answer|><|fim_prefix|><|fim_middle|><|fim_suffix|>If you use or build on this model, please retain attribution to Veyra AI.
Apache 2.0.
from transformers import AutoTokenizer, AutoModelForCausalLMimport torchrepo = "veyra-ai/Veyra-30M-Base"tokenizer = AutoTokenizer.from_pretrained(repo, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained(repo, trust_remote_code=True, dtype=torch.float32)model.eval()prompt = "Photosynthesis is the process by which"input_ids = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt")with torch.no_grad(): out = model.generate( input_ids, do_sample=True, temperature=0.5, top_k=30, repetition_penalty=1.15, no_repeat_ngram_size=2, max_new_tokens=80, )print(tokenizer.decode(out[0], skip_special_tokens=True))<|bos|><|eos|><|pad|><|unk|><|im_start|><|im_end|><|tool_call|><|tool_result|><|context|><|reasoning|><|end_reasoning|><|answer|><|fim_prefix|><|fim_middle|><|fim_suffix|>