praveenkumarpranjal commited on
Commit
ee3d460
·
verified ·
1 Parent(s): 6454e6c

Update links after username change

Browse files
Files changed (1) hide show
  1. README.md +6 -6
README.md CHANGED
@@ -12,7 +12,7 @@ I build practical, inspectable tools around open models: adapters, quantization
12
 
13
  ## Featured work
14
 
15
- ### [PathPack-Q · LFM2.5-2.6B training-free quantization](https://huggingface.co/alpha7845/LFM2.5-2.6B-4bit-PathPack-Q)
16
 
17
  An architecture-specific post-training quantization experiment for LiquidAI's
18
  hybrid convolution/attention LLM. PathPack-Q uses exact gated-path channel
@@ -24,7 +24,7 @@ without training, text calibration data, extra parameters, or runtime operators.
24
  - **Identical 1,517,616,892-byte checkpoint** and 4.501 effective bits/weight
25
  - Includes the search algorithm, complete-path acceptance gate, rejected-layer evidence, checkpoint builder, and machine-readable evaluations
26
 
27
- ### [ScopeGuard · Qwen2.5-1.5B LoRA](https://huggingface.co/alpha7845/scopeguard-qwen2.5-1.5b-lora)
28
 
29
  A locally trained agent decision-layer adapter that turns natural-language requests into strict JSON risk and confirmation decisions before tools execute.
30
 
@@ -33,15 +33,15 @@ A locally trained agent decision-layer adapter that turns natural-language reque
33
  - 3.957M trainable parameters — only 0.256% of the 1.5B base model
34
  - Includes adapter weights, original dataset, deterministic generator, training config, baseline outputs, and per-example evaluation
35
 
36
- → [Explore the ScopeGuard dataset](https://huggingface.co/datasets/alpha7845/scopeguard-decisions)
37
 
38
- → [Open the complete benchmark explorer](https://huggingface.co/spaces/alpha7845/scopeguard-benchmark)
39
 
40
- ### [LoRA Lens](https://huggingface.co/spaces/alpha7845/lora-lens-demo)
41
 
42
  An in-browser audit tool for `adapter_config.json` files. It surfaces rank, alpha, scaling, target modules, reproducibility gaps, and conservative parameter-efficiency estimates without uploading weights or requiring an API key.
43
 
44
- → [Inspect the implementation](https://huggingface.co/alpha7845/lora-lens)
45
 
46
  ## Current lab
47
 
 
12
 
13
  ## Featured work
14
 
15
+ ### [PathPack-Q · LFM2.5-2.6B training-free quantization](https://huggingface.co/praveenkumarpranjal/LFM2.5-2.6B-4bit-PathPack-Q)
16
 
17
  An architecture-specific post-training quantization experiment for LiquidAI's
18
  hybrid convolution/attention LLM. PathPack-Q uses exact gated-path channel
 
24
  - **Identical 1,517,616,892-byte checkpoint** and 4.501 effective bits/weight
25
  - Includes the search algorithm, complete-path acceptance gate, rejected-layer evidence, checkpoint builder, and machine-readable evaluations
26
 
27
+ ### [ScopeGuard · Qwen2.5-1.5B LoRA](https://huggingface.co/praveenkumarpranjal/scopeguard-qwen2.5-1.5b-lora)
28
 
29
  A locally trained agent decision-layer adapter that turns natural-language requests into strict JSON risk and confirmation decisions before tools execute.
30
 
 
33
  - 3.957M trainable parameters — only 0.256% of the 1.5B base model
34
  - Includes adapter weights, original dataset, deterministic generator, training config, baseline outputs, and per-example evaluation
35
 
36
+ → [Explore the ScopeGuard dataset](https://huggingface.co/datasets/praveenkumarpranjal/scopeguard-decisions)
37
 
38
+ → [Open the complete benchmark explorer](https://huggingface.co/spaces/praveenkumarpranjal/scopeguard-benchmark)
39
 
40
+ ### [LoRA Lens](https://huggingface.co/spaces/praveenkumarpranjal/lora-lens-demo)
41
 
42
  An in-browser audit tool for `adapter_config.json` files. It surfaces rank, alpha, scaling, target modules, reproducibility gaps, and conservative parameter-efficiency estimates without uploading weights or requiring an API key.
43
 
44
+ → [Inspect the implementation](https://huggingface.co/praveenkumarpranjal/lora-lens)
45
 
46
  ## Current lab
47