Instructions to use beezu/Q3.5-BlueStar-v2-27B-mlx-bf16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use beezu/Q3.5-BlueStar-v2-27B-mlx-bf16 with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("beezu/Q3.5-BlueStar-v2-27B-mlx-bf16")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use beezu/Q3.5-BlueStar-v2-27B-mlx-bf16 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "beezu/Q3.5-BlueStar-v2-27B-mlx-bf16"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "beezu/Q3.5-BlueStar-v2-27B-mlx-bf16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use beezu/Q3.5-BlueStar-v2-27B-mlx-bf16 with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "beezu/Q3.5-BlueStar-v2-27B-mlx-bf16"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default beezu/Q3.5-BlueStar-v2-27B-mlx-bf16

Run Hermes

hermes

MLX LM

How to use beezu/Q3.5-BlueStar-v2-27B-mlx-bf16 with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "beezu/Q3.5-BlueStar-v2-27B-mlx-bf16"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "beezu/Q3.5-BlueStar-v2-27B-mlx-bf16"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "beezu/Q3.5-BlueStar-v2-27B-mlx-bf16",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

beezu/Q3.5-BlueStar-v2-27B-mlx-bf16

This model beezu/Q3.5-BlueStar-v2-27B-mlx-bf16 was converted to MLX format from zerofata/Q3.5-BlueStar-v2-27B using mlx-lm version 0.31.2.

Original Model Card

BlueStar

BlueStar v2

Qwen3.5 27B

01 Overview

Designed for RP and writing tasks.

Feels like a good improvement on v1. This version aims to fix the rep and improve the intelligence while keeping the creativity.

Non thinking and thinking are both supported. If you want to use thinking, it is required to prefill the <think>\n as that is how it was trained.

02 SillyTavern Settings

Recommended Roleplay Format

ActionsIn plaintext

Dialogue"In quotes"

Thoughts*In asterisks*

Recommended Samplers

Temp0.8 - 1.0

MinP0.05 - 0.075

Instruct

ChatML - Think

ChatML - NoThink

03 Quantizations

GGUF

iMatrix

04 Creation Process

Creation Process: SFT

SFT on approx 27 million tokens.

I've confirmed the repetition coming from the RP datasets. Despite the extensive filtering, human editing, rewriting and deduping. Compared to other types of data like chat and writing, RP is just somewhat repetitive in nature. One idea to fix this is to just not use the RP datasets, or use less of them. This does seem to *sort of* work, but the model performs noticably worse at RP as a result. Which makes sense, given that's the entire idea of having RP data to begin with.

The current solution I'm testing is using custom loss masking with the RP datasets. Most common phrases of slop are masked out, so the model doesn't get rewarded for learning these patterns. Overused words within a conversation also get masked out in later turns.

It... seems to have worked? Repetition from my testing is greatly reduced after a few hours of using the model. It can still latch onto phrases, but I've seen much less verbatim repetition.

Trained using Axolotl.

Axolotl Config

SFT (4×H200)

base_model: Qwen/Qwen3.5-27B
 
plugins:
  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
strict: false
 
datasets:
  - path: ./data/bluestar_v2_sft_3_all_rp_attempt_masked_20260318_075236.jsonl
 
val_set_size: 0.02
output_dir: ./Qwen3.5-27B-v2-SFT-5
 
sequence_len: 10756
sample_packing: true
 
load_in_8bit: true
adapter: lora
lora_r: 128
lora_alpha: 128
peft_use_rslora: true
lora_target_modules:
  - q_proj
  - k_proj
  - v_proj
  - o_proj
  - down_proj
  - up_proj
  # Uncomment below to also target the linear attention projections.
  # These use separate in_proj_qkv / in_proj_z / out_proj (Qwen3.5-specific).
  - linear_attn.in_proj_qkv
  - linear_attn.in_proj_z
  - linear_attn.out_proj
 
wandb_project: Qwen3.5-27B-SFT
wandb_name: Qwen3.5-27B-v2-SFT-5
 
gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 2
optimizer: adamw_torch_8bit
lr_scheduler: cosine
learning_rate: 1.2e-5
weight_decay: 0.01
warmup_ratio: 0.05
 
bf16: auto
tf32: true
 
resume_from_checkpoint:
logging_steps: 1
flash_attention: true
 
evals_per_epoch: 4
saves_per_epoch: 4
special_tokens:
 
fsdp_config:
  fsdp_version: 2
  offload_params: false
  cpu_ram_efficient_loading: false
  auto_wrap_policy: TRANSFORMER_BASED_WRAP
  transformer_layer_cls_to_wrap: Qwen3_5DecoderLayer
  state_dict_type: FULL_STATE_DICT
  sharding_strategy: FULL_SHARD
  reshard_after_forward: true
  activation_checkpointing: true