Instructions to use yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp with Transformers:

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

pipe = pipeline("text-generation", model="yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp")
model = AutoModelForCausalLM.from_pretrained("yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp

SGLang

How to use yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp 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 "yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp" \
    --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": "yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp",
		"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 "yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp" \
        --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": "yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp with Docker Model Runner:
```
docker model run hf.co/yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp
```

C5 ties_sum rho=0.5 lambda=0.7 — merge ablation (NOT the recommended checkpoint)

This is a negative result published for completeness. It performs worse than yungisimon/Qwen2.5-14B-ties-merge-nqa-msq-bcp, which is the checkpoint to use. This one exists to document what was tried and ruled out.

Merges three 14B MEMORY models — each fine-tuned on a different corpus (NarrativeQA, MuSiQue, BrowseComp-Plus) — into one set of weights, entirely in parameter space.

Recipe

python main.py \
  --models <nqa> <msq> <bcp> \
  --base Qwen/Qwen2.5-14B-Instruct \
  --method ties_sum --weight 0.7 0.7 0.7 --density 0.5 0.5 0.5 --raw-weights --anchor-base

Task vectors are computed against the pretrained base, tau_i = phi_i - phi_0.


alpha = <D_merged, tau_i>/\|\|tau_i\|\|^2 (nqa / msq / bcp)	0.72 / 0.74 / 0.77
\|\|D_merged\|\| / \|\|phi_0\|\|	8.0%
single-corpus fine-tune, for reference	4.8–7.0%

alpha = 1.0 would mean a corpus's fine-tuning survived the merge at full strength.

Measured result

Single-turn knowledge recall: the MEMORY model is asked corpus questions directly, with no multi-turn protocol, judged against gold answers by Qwen2.5-32B-Instruct. n=80 per corpus, restricted to the subsets each model was actually trained on.

model	BrowseComp-Plus	MuSiQue
this checkpoint	7.5%	18.8%
`yungisimon/Qwen2.5-14B-ties-merge-nqa-msq-bcp` (recommended)	16.2%	25.0%
per-corpus specialist (ceiling)	~37%	~36%
untrained Qwen2.5-14B-Instruct (floor)	0%	~2.5%

The reference figures are from the same probe run as this checkpoint, because cross-run noise is about +/-5pp at n=80 (the reference measured 21.2 / 16.2 / 25.0 on BCP across three runs). Differences smaller than ~5pp are not resolvable at this sample size.

The decisive experiment: displacement matched to the reference (8.0% vs 7.9%) with sign-consensus filtering intact, but higher task-vector retention. It loses, which is what rules out retention as the objective.

Why these all failed

The three task vectors are near-orthogonal in aggregate (pairwise cosine 0.002–0.038), which suggests summing them rather than averaging, since orthogonal updates should not interfere. That reasoning is wrong: 75.4% of coordinates carry a sign conflict between the three vectors. Global orthogonality does not imply coordinate-wise compatibility. TIES's magnitude trimming and sign-consensus filtering are therefore doing necessary conflict resolution, not merely attenuating the task vectors — so recipes that preserve more of each task vector (higher alpha) consistently do worse. Six candidates across four families (pure sum, sparse sum, sub-0.3 sparsity, non-uniform weighting) all lost to plain TIES rho=0.3 with uniform averaging.

Caveats

The probe asks the full question in a single turn, whereas the MEMORY model is trained on decomposed, self-contained reflection pairs and is used at inference on decomposed sub-queries. Absolute numbers are therefore well below published protocol accuracy and are not comparable to it; the probe is a relative screen of knowledge accessibility.
Judging is a single greedy pass of one local model, not human-validated.
A NarrativeQA column was collected but discarded: the NQA specialist scored 0% on its own corpus, so that probe measures prompt-style mismatch rather than knowledge.

License

Inherits the license of Qwen/Qwen2.5-14B-Instruct (Apache 2.0).

Downloads last month: 24

Safetensors

Model size

15B params

Tensor type

BF16

Model tree for yungisimon/Qwen2.5-14B-merge-tiessum-r0.5-l0.7-nqa-msq-bcp

1010stonely/Qwen2.5-14B_BCP300_step1-7_N_neg_updated_batchsize1024_trngepoch3

Glow-AI/QWEN2.5-14b_instruct-MSQ1000_step1-5_with_N_neg_batchsize256_best_epoch

Glow-AI/QWEN2.5-14b_instruct-NQA10_step1-5_batchsize512_best_epoch

Qwen/Qwen2.5-14B-Instruct

Merge model

this model