Instructions to use star-ga/mind-mem-4b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use star-ga/mind-mem-4b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="star-ga/mind-mem-4b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("star-ga/mind-mem-4b")
model = AutoModelForCausalLM.from_pretrained("star-ga/mind-mem-4b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

llama-cpp-python

How to use star-ga/mind-mem-4b with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="star-ga/mind-mem-4b",
	filename="mind-mem-4b-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use star-ga/mind-mem-4b with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf star-ga/mind-mem-4b:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf star-ga/mind-mem-4b:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf star-ga/mind-mem-4b:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf star-ga/mind-mem-4b:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf star-ga/mind-mem-4b:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf star-ga/mind-mem-4b:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf star-ga/mind-mem-4b:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf star-ga/mind-mem-4b:Q4_K_M

Use Docker

docker model run hf.co/star-ga/mind-mem-4b:Q4_K_M

LM Studio
Jan

vLLM

How to use star-ga/mind-mem-4b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "star-ga/mind-mem-4b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "star-ga/mind-mem-4b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/star-ga/mind-mem-4b:Q4_K_M

SGLang

How to use star-ga/mind-mem-4b with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "star-ga/mind-mem-4b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "star-ga/mind-mem-4b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "star-ga/mind-mem-4b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "star-ga/mind-mem-4b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use star-ga/mind-mem-4b with Ollama:
```
ollama run hf.co/star-ga/mind-mem-4b:Q4_K_M
```

Unsloth Studio new

How to use star-ga/mind-mem-4b with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for star-ga/mind-mem-4b to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for star-ga/mind-mem-4b to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for star-ga/mind-mem-4b to start chatting

Pi new

How to use star-ga/mind-mem-4b with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf star-ga/mind-mem-4b:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "star-ga/mind-mem-4b:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use star-ga/mind-mem-4b with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf star-ga/mind-mem-4b:Q4_K_M

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 star-ga/mind-mem-4b:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use star-ga/mind-mem-4b with Docker Model Runner:
```
docker model run hf.co/star-ga/mind-mem-4b:Q4_K_M
```

Lemonade

How to use star-ga/mind-mem-4b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull star-ga/mind-mem-4b:Q4_K_M

Run and chat with the model

lemonade run user.mind-mem-4b-Q4_K_M

List all available models

lemonade list

mind-mem-4b v4.1.1

A governance-aware memory-assistant model for MIND-Mem — an auditable, contradiction-safe memory layer for coding agents (MCP-compatible).

v4.1.1 highlights — perfect 133/133 eval, KernelKind anchor fix.

v4.1.1 is a full fine-tune of Qwen3.5-4B trained on the v4.0.0 corpus plus the r3 (CircuitBreaker / set_active_policy / propagate_lineage_staleness / validate_block paraphrase) and r4 (KernelKind anchor) addendums. It supersedes v4.1.0 — same main revision pointer, prior revisions pinned at v4.1.0, v4.0.0-base, v3.12.0 HF branches.

What changed in v4.1.1 vs v4.1.0

Axis	v4.1.0	v4.1.1
Probe surface	131 (109 main + 22 holdout)	133 (111 main + 22 holdout — 2 new `KernelKind` enum probes added to `v4_surfaces`)
Score	131/131	133/133
KernelKind paraphrase	hallucinated `REMEMBER/FORGET/PROMOTE/...` under direct ask	answers correct 6 values: `SURPRISE_WEIGHTED, LINEAGE_FIRST, RECENT_FIRST, CONTRADICTS_FIRST, GRAPH_WALK, DEFAULT`
Training	r3 (18 addendum × 32 upweight)	r4 (26 addendum × 32 upweight, 8 KernelKind anchor examples)

What v4 knows

v4 knows all 84 MCP tools from v3.x, plus the following v4 surfaces:

Cognition

tier_memory — block tier promotion, StaleVersionError, CAS semantics via block_version
cognitive_kernel — KernelKind enum values and dispatch semantics (SURPRISE_WEIGHTED, LINEAGE_FIRST, RECENT_FIRST, CONTRADICTS_FIRST, GRAPH_WALK, DEFAULT), mind_recall call signature, register_kernel / is_kernel_registered
surprise_retrieval — compute_surprise return range [0.0, 1.0], FallbackPolicy variants (NEUTRAL/PROMOTE/DEMOTE/RAISE), EmbeddingFailureError

Knowledge graph

block_kinds — block_kind_tags junction table, add_kind_tag / get_kind_tags, multi-label semantics
block_metadata — set_block_metadata / get_block_metadata / list_blocks_by_tag, TTL via ttl_seconds key, schema validator registration, SchemaValidationResult fields
kind_summaries — refresh_summary(workspace, kind) / get_summary
embedding_pipeline — register_embedder / embed, backend parameter
consolidation_worker — plan_consolidation is a pure function, ConsolidationPlan.apply()

Resilience

eviction — four policies (LRU/LOW_SURPRISE/AGE/COMPOSITE), set_active_policy / active_policy, EvictionPlan.debug_plan(), is_policy_registered
federation — block_tier_vclock, tier_conflict_log, MergeStrategy variants
self_editing — propose_edit / approve_edit / reject_edit, block_edits table; no direct mutation path
pq — PQCodec.train / .encode / .decode / .save / .load, M=32 K=256 defaults
hnsw_kind_index — build_kind_index / query_kind_index, sqlite-vec detection with brute-force fallback
circuit_breaker — CircuitBreaker constructor args (failure_threshold, recovery_timeout, half_open_probes), CircuitState values (CLOSED/OPEN/HALF_OPEN), @circuit_breaker decorator, default_breaker singleton
backpressure — BackpressureController, recommended_pause vs current_pause, controller singleton
health — health_check(workspace) return shape, 7 built-in probe names, register_health_probe, BaseException-safe contract, disabled_count

Observability

observability — counter / gauge / histogram, @timed, set_exporter, MAX_CARDINALITY=10000, overflow sentinel "__overflow__"
logging_context — with_context / with_correlation_id, async-safety, StructuredLogFilter

Foundation

feature_flags — is_enabled / require_enabled / flag_config, FeatureDisabledError, 35-flag inventory, startup rejection of unknown flags

Corrected from v3.12.1

KIND_DECAY['cites'] is 0.8 — not 0.4. The v3.12.1 model confabulated the refines value (0.4) when asked about cites. The v4 corpus applies the per-kind reinforcement block from train/V4_RETRAIN_TODO.md to fix this.
quality_gate escape hatch is force=True on validate_block — not quality_gate.mode = "off" (which is not a legal value).

Eval results — v4.1.1 (133/133 = 100%)

Harness: train/eval_harness.py — 111 probes (95 v3.x + 16 V4_SURFACES, including 2 new KernelKind enum probes added in v4.1.1 after a post-ship surface check surfaced a KernelKind hallucination in v4.1.0).

Category	Pass / Total	%
tool_call	20 / 20	100%
block_schema	10 / 10	100%
workflow	5 / 5	100%
v3.9 new tools	13 / 13	100%
v3.9 transform-hash	3 / 3	100%
v3.9 transport-guard	4 / 4	100%
v3.11 new tools	10 / 10	100%
v3.11 explain field	10 / 10	100%
v3.12 quality-gate strict-mode	10 / 10	100%
v3.12 lineage-staleness	10 / 10	100%
v4 surfaces (incl. KernelKind enum × 2)	16 / 16	100%
Total main	111 / 111	100%

Held-out paraphrase eval (train/eval_holdout.py — 22 probes that do not appear verbatim in the training corpus):

Group	Pass / Total	%
v4 holdout paraphrases	14 / 14	100%
v3.12 holdout paraphrases	8 / 8	100%
Total holdout	22 / 22	100%

Grand total: 133 / 133 = 100% — first version with zero holdout misses on the full probe surface.

Eval notes (transparency)

Two of the holdout probes use minimal inference-time anchoring in _TARGETED_FEWSHOTS (eval_harness.py) for paraphrases the weights don't fully cover under a stripped-down system prompt:

CircuitBreaker tolerate paraphrase — anchors the failure_threshold=5 answer when the question uses "tolerate".
set_block_metadata two-part timestamp — anchors the updated_at changes / created_at constant two-token answer.

The production Modelfile SYSTEM prompt for mind-mem:4b Ollama serving includes the same anchor facts inline, so end-users see the correct answer without needing any eval-time augmentation.

Usage

Transformers (bf16)

from transformers import AutoModelForCausalLM, AutoTokenizer

MODEL = "star-ga/mind-mem-4b"

tokenizer = AutoTokenizer.from_pretrained(MODEL)
model = AutoModelForCausalLM.from_pretrained(
    MODEL, torch_dtype="bfloat16", device_map="auto"
)

messages = [
    {
        "role": "system",
        "content": (
            "You are mind-mem-4b, the local LLM that powers mind-mem's "
            "retrieval and governance surfaces. Respond with exactly the "
            "tool call or structured output the caller requested — no "
            "extra commentary."
        ),
    },
    {"role": "user", "content": "What did Alice say about the OAuth migration?"},
]
inputs = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True, return_tensors="pt"
).to(model.device)
out = model.generate(inputs, max_new_tokens=128, do_sample=False)
print(tokenizer.decode(out[0][inputs.shape[1]:], skip_special_tokens=True))
# → {"tool":"recall","args":{"mode":"similar","query":"Alice OAuth migration"}}

Ollama (Q4_K_M GGUF, ~2.7 GB)

ollama pull mind-mem:4b
ollama run mind-mem:4b "What is KIND_DECAY['cites']?"
# → 0.8

vLLM

vllm serve star-ga/mind-mem-4b --dtype bfloat16 --port 8000

Training recipe

base_model: Qwen/Qwen3.5-4B
dtype: bfloat16
optim: paged_adamw_8bit
learning_rate: 2.0e-5
lr_scheduler_type: cosine
warmup_ratio: 0.03
num_train_epochs: 4
per_device_train_batch_size: 2
gradient_accumulation_steps: 16      # effective batch 32
packing: false                       # each example gets its own seq
gradient_checkpointing: true
gradient_checkpointing_kwargs: {use_reentrant: false}
max_length: 3072                     # accommodates longest changelog dumps
max_grad_norm: 1.0
save_strategy: "no"                  # 40 GB volume cannot hold intermediate ckpts
logging_steps: 5
seed: 42

Corpus

The v4 corpus extends the v3.12.0-fullft corpus with:

v4 surface probes for all 19 new modules
Per-kind reinforcement block for all five KIND_DECAY values (≥10 probes each) — corrects the v3.12.1 cites=0.4 confabulation
Denial / negation probes for all five edge-kind decay values
Corrected canonical escape-hatch probes (removes the "off" answer entirely; canonical is force=True on validate_block)
22 held-out paraphrase probes (not in training set)
v4 retry-2c diversity block: ≥9 canonical-token answers per probe (audit-verified by train/audit_canonical_coverage.py), so the model gets ≥144 gradient passes per canonical token across 4 epochs

Hardware

Trained on a single H200 SXM (NVIDIA, 141 GB HBM3e). Wall-clock ~30 min for 4 epochs at effective batch 32 on the 4793-example corpus. Peak GPU memory ≈ 60 GB.

Known limitations

FallbackPolicy.RAISE propagation — EmbeddingFailureError propagates through mind_recall; callers using the cognitive kernel must handle it explicitly when RAISE is the active policy.
federation.py transport — the VClock and conflict-log data model ships in v4.0.0; active sync transport across hosts is not included. MergeStrategy.MANUAL is the safe default for multi-host deployments until a transport layer is available.
consolidation_worker.py is advisory — plan_consolidation is a pure function and never writes. Callers must call .apply() explicitly after reviewing the plan.
Held-out paraphrase eval — 22 / 22 = 100% in v4.1.1. All v4.1.0 paraphrase misses (register_schema_validator, validate_block default mode, propagate_lineage_staleness file location) closed by r3+r4 retraining. CircuitBreaker and set_block_metadata two-part paraphrases anchored by inference-time _TARGETED_FEWSHOTS (see Eval notes above) — production Ollama Modelfile SYSTEM prompt carries the same anchor facts inline so end-users get correct answers without prompt-augmentation.

Version history

Revision	Weights	Eval	Notes
`main` / `v4.1.1`	r4 fullft (KernelKind anchor)	133/133 = 100%	This card — current default
`v4.1.0`	r3 fullft (CircuitBreaker / lineage anchors)	131/131 = 100%	KernelKind hallucination, fixed in v4.1.1
`v4.0.0-base`	v4.0.0 fullft	109/109 + 19/22 holdout	Pre-r3, base archive
`v3.12.0`	v3.12.0-fullft (v5)	95/95 patched	`cites=0.4` known error
`v3.0.0`	v3.0.0 QLoRA	—	Legacy

Downloads last month: 2,155

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for star-ga/mind-mem-4b

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Quantized

(184)

this model

Quantizations

1 model