Hugging Face's logo

hharsha
/
agentic-github-tagger

Text Generation
Transformers
Safetensors
PEFT
English
t5
text2text-generation
agents
rag
llmops
lora
text-generation-inference
Model card Files
xet
 Community
hharsha commited on
Commit
0be48af
·
verified ·
1 Parent(s): 3a2b839

Retrain 3 epochs + clearer sample outputs

Browse files
Files changed (3) hide show
  1. README.md +20 -31
  2. model.safetensors +1 -1
  3. training_metadata.json +15 -3
README.md CHANGED
@@ -10,8 +10,8 @@ tags:
10
  - agents
11
  - rag
12
  - llmops
13
- - text2text-generation
14
  - lora
 
15
  datasets:
16
  - hharsha/agentic-github-meta
17
  pipeline_tag: text2text-generation
@@ -25,24 +25,11 @@ widget:
25
 
26
  Lightweight **text2text tag generator** for agentic AI / RAG / LLMOps GitHub-style
27
  descriptions. Fine-tuned from [`google-t5/t5-small`](https://huggingface.co/google-t5/t5-small)
28
- with **PEFT LoRA** on [`hharsha/agentic-github-meta`](https://huggingface.co/datasets/hharsha/agentic-github-meta),
29
- then **merged** into full small weights so `transformers` pipelines work out of the box on CPU.
30
-
31
- > This is a **~60M-param T5-small** tagger — **not** a 7B chat demo. Free Hub + CPU friendly.
32
-
33
- ## LoRA config (training)
34
 
35
- | | |
36
- |---|---|
37
- | Base | `google-t5/t5-small` |
38
- | Method | PEFT LoRA then merge |
39
- | r | 16 |
40
- | alpha | 32 |
41
- | dropout | 0.05 |
42
- | target_modules | `q`, `v` |
43
- | Epochs | 2 (CPU) |
44
- | Batch size | 8 |
45
- | Dataset | [`hharsha/agentic-github-meta`](https://huggingface.co/datasets/hharsha/agentic-github-meta) |
46
 
47
  ## Usage
48
 
@@ -52,23 +39,25 @@ pipe = pipeline("text2text-generation", model="hharsha/agentic-github-tagger")
52
  print(pipe("multi-agent platform with RAG, MCP, and observability")[0]["generated_text"])
53
  ```
54
 
55
- Sample output from this training run:
56
 
57
  ```
58
- observable, multi-agent, rAG, mCP, observability, et-
59
  ```
60
 
61
- ## Links
62
-
63
- - Dataset: [`hharsha/agentic-github-meta`](https://huggingface.co/datasets/hharsha/agentic-github-meta)
64
- - Showcase: [`hharsha/agentic-systems-showcase`](https://huggingface.co/datasets/hharsha/agentic-systems-showcase)
65
- - Studio: [https://agentic-systems-studio.com](https://agentic-systems-studio.com)
66
- - GitHub: [https://github.com/hharsha98](https://github.com/hharsha98)
67
 
68
- ## Intended use
69
-
70
- Auto-suggest comma-separated tags for agentic / RAG / LLMOps project listings and search demos.
 
 
 
 
71
 
72
- ## Limitations
73
 
74
- Small model; tags can be generic or incomplete. Not for safety-critical labeling.
 
 
 
 
10
  - agents
11
  - rag
12
  - llmops
 
13
  - lora
14
+ - text2text-generation
15
  datasets:
16
  - hharsha/agentic-github-meta
17
  pipeline_tag: text2text-generation
 
25
 
26
  Lightweight **text2text tag generator** for agentic AI / RAG / LLMOps GitHub-style
27
  descriptions. Fine-tuned from [`google-t5/t5-small`](https://huggingface.co/google-t5/t5-small)
28
+ with **PEFT LoRA** (r=16, alpha=32, dropout=0.05, target_modules q/v) on
29
+ [`hharsha/agentic-github-meta`](https://huggingface.co/datasets/hharsha/agentic-github-meta),
30
+ then **merged** so `transformers` pipelines work on free CPU.
 
 
 
31
 
32
+ > ~60M-param T5-small tagger — **not** a 7B chat demo.
 
 
 
 
 
 
 
 
 
 
33
 
34
  ## Usage
35
 
 
39
  print(pipe("multi-agent platform with RAG, MCP, and observability")[0]["generated_text"])
40
  ```
41
 
42
+ Sample:
43
 
44
  ```
45
+ Multi-agents, multi-agent, observability, observability, observability, rag, mCP, observability, multiagents, multi-agents, observability, observability, observability, observability,
46
  ```
47
 
48
+ ## Training
 
 
 
 
 
49
 
50
+ | | |
51
+ |---|---|
52
+ | Base | `google-t5/t5-small` |
53
+ | LoRA | r=16, alpha=32, dropout=0.05, modules q/v |
54
+ | Epochs | 3 (CPU) |
55
+ | Batch | 8 |
56
+ | Dataset | [`hharsha/agentic-github-meta`](https://huggingface.co/datasets/hharsha/agentic-github-meta) |
57
 
58
+ ## Links
59
 
60
+ - Dataset: https://huggingface.co/datasets/hharsha/agentic-github-meta
61
+ - Showcase: https://huggingface.co/datasets/hharsha/agentic-systems-showcase
62
+ - Studio: https://agentic-systems-studio.com
63
+ - GitHub: https://github.com/hharsha98
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a3b4dda9b0e5d4c45c9ee0a9d83e35a8b6c3f3ceab2be3fa176c1020b0051d8c
3
  size 242041896
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a64832c58df0fe26aeadbb04445fecc46f6731974c76190d9a9d7aad3b755c98
3
  size 242041896
training_metadata.json CHANGED
@@ -9,10 +9,22 @@
9
  "v"
10
  ]
11
  },
12
- "epochs": 2,
13
  "batch_size": 8,
14
  "train_rows": 600,
15
- "sample_input": "multi-agent platform with RAG, MCP, and observability",
16
- "sample_output": "observable, multi-agent, rAG, mCP, observability, et-",
 
 
 
 
 
 
 
 
 
 
 
 
17
  "merged": true
18
  }
 
9
  "v"
10
  ]
11
  },
12
+ "epochs": 3,
13
  "batch_size": 8,
14
  "train_rows": 600,
15
+ "samples": [
16
+ {
17
+ "input": "multi-agent platform with RAG, MCP, and observability",
18
+ "output": "Multi-agents, multi-agent, observability, observability, observability, rag, mCP, observability, multiagents, multi-agents, observability, observability, observability, observability, "
19
+ },
20
+ {
21
+ "input": "looking for RAG hybrid search recall@k and reranking",
22
+ "output": "reranking, recall@k, reranking,"
23
+ },
24
+ {
25
+ "input": "FastAPI Next.js agent backend with Docker Compose",
26
+ "output": "FastAPI, Next.js, backend, agent, docker, comma"
27
+ }
28
+ ],
29
  "merged": true
30
  }