Text Generation
MLX
Safetensors
qwen2
1.5b
alloy-backfilled
android
apple-silicon
attested
chain-of-custody
chinese
compacted
consumer-gpu
cryptographically-verified
edge-inference
efficient
embedded
english
forge-alloy
general
general-purpose
head-pruning
iphone
llama-cpp
lm-studio
local-inference
macbook
mobile
multilingual
ollama
on-device
optimized
pruned
qwen
qwen2.5
raspberry-pi
reproducible
versatile
Instructions to use continuum-ai/qwen2.5-1.5b-general-forged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use continuum-ai/qwen2.5-1.5b-general-forged with MLX:
# Make sure mlx-lm is installed # pip install --upgrade mlx-lm # if on a CUDA device, also pip install mlx[cuda] # Generate text with mlx-lm from mlx_lm import load, generate model, tokenizer = load("continuum-ai/qwen2.5-1.5b-general-forged") prompt = "Once upon a time in" text = generate(model, tokenizer, prompt=prompt, verbose=True) - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- MLX LM
How to use continuum-ai/qwen2.5-1.5b-general-forged with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Generate some text mlx_lm.generate --model "continuum-ai/qwen2.5-1.5b-general-forged" --prompt "Once upon a time"
Correct qwen2.5-1.5b-general-forged.alloy.json pass@1 to canonical evalplus convention (v1.0.0)
Browse files
qwen2.5-1.5b-general-forged.alloy.json
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "qwen2.5-1.5b-general-forged",
|
| 3 |
+
"version": "1.0.0",
|
| 4 |
+
"description": "Forged Qwen2.5-1.5B for general domain via per-layer head pruning + LoRA recovery training. This alloy was retroactively synthesized from forging_results.json on 2026-04-08 \u2014 the forge run itself executed at 2026-03-27T09:01:22-05:00. The published model weights are unchanged; this alloy adds the missing forge-alloy provenance envelope so the artifact participates in the chain-of-custody system.",
|
| 5 |
+
"author": "continuum-ai",
|
| 6 |
+
"tags": [
|
| 7 |
+
"general",
|
| 8 |
+
"forged",
|
| 9 |
+
"experiential-plasticity",
|
| 10 |
+
"forge-alloy",
|
| 11 |
+
"alloy-backfilled"
|
| 12 |
+
],
|
| 13 |
+
"license": "apache-2.0",
|
| 14 |
+
"source": {
|
| 15 |
+
"baseModel": "Qwen/Qwen2.5-1.5B",
|
| 16 |
+
"architecture": "qwen2"
|
| 17 |
+
},
|
| 18 |
+
"stages": [
|
| 19 |
+
{
|
| 20 |
+
"type": "prune",
|
| 21 |
+
"strategy": "magnitude",
|
| 22 |
+
"level": 0.3,
|
| 23 |
+
"notes": "Legacy strategy name from forging_results.json: 'combined'. Mapped to canonical 'magnitude' for the alloy schema. The actual forge run used the legacy code path; this alloy is a retroactive provenance record, not a re-execution recipe."
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"type": "train",
|
| 27 |
+
"domain": "general",
|
| 28 |
+
"steps": 1000,
|
| 29 |
+
"learningRate": "2e-4",
|
| 30 |
+
"dataset": "wikitext-2"
|
| 31 |
+
}
|
| 32 |
+
],
|
| 33 |
+
"cycles": 3,
|
| 34 |
+
"results": {
|
| 35 |
+
"completedAt": "2026-03-27T09:01:22-05:00",
|
| 36 |
+
"baselinePerplexity": 2.5,
|
| 37 |
+
"finalPerplexity": 2.44,
|
| 38 |
+
"improvementPct": 2.4,
|
| 39 |
+
"benchmarks": [
|
| 40 |
+
{
|
| 41 |
+
"name": "perplexity",
|
| 42 |
+
"metrics": {
|
| 43 |
+
"baseline": 2.5,
|
| 44 |
+
"final": 2.44,
|
| 45 |
+
"improvement_pct": 2.4,
|
| 46 |
+
"dataset": "wikitext-2"
|
| 47 |
+
}
|
| 48 |
+
}
|
| 49 |
+
],
|
| 50 |
+
"hardwareVerified": [
|
| 51 |
+
{
|
| 52 |
+
"device": "MacBook Air 8GB",
|
| 53 |
+
"format": "fp16",
|
| 54 |
+
"verified": false
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"device": "MacBook Pro 16GB",
|
| 58 |
+
"format": "fp16",
|
| 59 |
+
"verified": true
|
| 60 |
+
}
|
| 61 |
+
],
|
| 62 |
+
"samples": [],
|
| 63 |
+
"integrity": {
|
| 64 |
+
"trustLevel": "self-attested",
|
| 65 |
+
"code": {
|
| 66 |
+
"runner": "sentinel-ai/forge_model (legacy pre-\u00a74.1.3.1 path)",
|
| 67 |
+
"version": "2.x",
|
| 68 |
+
"binaryHash": "sha256:legacy-pre-alloy-schema"
|
| 69 |
+
},
|
| 70 |
+
"modelHash": "sha256:21ca799dd3ee2f73526c9422b69bc9f931e8a22be34f21e1bc537a4df86a998a",
|
| 71 |
+
"fileHashes": [
|
| 72 |
+
{
|
| 73 |
+
"filename": "model.safetensors",
|
| 74 |
+
"sha256": "84d005e8a74c3feda47abd286622e4b95d37f055587593a9423c659a2a550138",
|
| 75 |
+
"size": 3087467144
|
| 76 |
+
}
|
| 77 |
+
],
|
| 78 |
+
"datasets": [
|
| 79 |
+
{
|
| 80 |
+
"name": "wikitext-2",
|
| 81 |
+
"hash": "sha256:not-pinned-legacy"
|
| 82 |
+
}
|
| 83 |
+
],
|
| 84 |
+
"attestedAt": "2026-03-27T09:01:22-05:00"
|
| 85 |
+
}
|
| 86 |
+
}
|
| 87 |
+
}
|