GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
HYU-NLP-EVA2
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
HYU-NLP-EVA2
Model Tree
Qwen/Qwen2.5-7B-Instruct
Input Modalities
Output Modalities
Supported Functionality
Inference export from 261019_ruscarl_baseline_rar_medicine, displayed canonically as
20261020_rar-medicine_static_qwen2.5-7b_raw4500_impl-ruscarl_s1. This repository contains the checkpoint saved at training step
50; the run has checkpoints through step 50 only.
Qwen/Qwen2.5-7B-Instruct at revision a09a35458c702b33eeacc393d103063234e8bc28.Qwen/Qwen3-32B at revision 9216db5781bf21249d130ec9da846c4624c16137, non-thinking,
criterion-level grading.HYU-NLP-EVA2/rar-medicine-ruscarl-test500-seed42,
selected with seed 42. The trainer data seed was the implicit default value 1;
these seed values describe different operations..pt shards are
excluded from the Hub payload. All local source checkpoints are retained.python
from transformers import AutoModelForCausalLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("HYU-NLP-EVA2/20261020_rar-medicine_static_qwen2.5-7b_raw4500_impl-ruscarl_s1_step-050")model = AutoModelForCausalLM.from_pretrained("HYU-NLP-EVA2/20261020_rar-medicine_static_qwen2.5-7b_raw4500_impl-ruscarl_s1_step-050", torch_dtype="bfloat16")