You can deploy this model and start earning money today!

osieosie/
tmax-qwen36-27b-tmax-sft-full-20260513-epoch1

Name: tmax-qwen36-27b-tmax-sft-full-20260513-epoch1
Availability: InStock

intermediate-checkpoint

hamishivi/Qwen3.6-27B

finetune

apache-2.0

Views

No views yet

tmax-qwen36-27b-tmax-sft-full-20260513-epoch1

Intermediate SFT checkpoint of hamishivi/Qwen3.6-27B produced by open-instruct.

This snapshot is the model state at the end of epoch 1 (after one full pass over the training set), exported from a DeepSpeed ZeRO-2 sharded checkpoint via scripts/convert_zero_checkpoint_to_hf.py.

Source checkpoint on disk: output/sft_qwen36_27b_tmax_full_0513_8node/epoch_0
Conversion: ZeRO-2 sharded BF16 → consolidated FP32 → cast to BF16 → HF safetensors
Tokenizer + chat template copied from hamishivi/Qwen3.6-27B

Use it like any HF causal-LM model:

python

from transformers import AutoModelForCausalLM, AutoTokenizer

tok = AutoTokenizer.from_pretrained("osieosie/tmax-qwen36-27b-tmax-sft-full-20260513-epoch1")
model = AutoModelForCausalLM.from_pretrained("osieosie/tmax-qwen36-27b-tmax-sft-full-20260513-epoch1", torch_dtype="bfloat16")