Spaces:

tusarway
/

rag-backend

Sleeping

App Files Files Community

rag-backend / model /loader.py

Commit History

Mar 6

05:14
fix: update model

imtrt004

ab16882
05:02
fix: remove exAI

imtrt004

6780118
04:45
fix: still error

imtrt004

ac40983
04:42
fix: still error

imtrt004

666afe0
04:27
fix: check_model_input issue

imtrt004

bd1ed9f
04:17
fix: update req

imtrt004

3e3b4c9
03:59
fix: update req

imtrt004

01693a1
03:48
fix: update req and docker

imtrt004

c227cde
03:32
fix: update base model

imtrt004

11d16b6
03:18
fix: exAI ropeparam

imtrt004

5f8085e

Mar 5

23:13
fix: update dependency

imtrt004

9f311d9

Mar 4

04:15
fix: update backend lib with log

imtrt004

2aa0b72
04:00
fix: update model lists

imtrt004

7997082
02:41
feat: select model from admin panel

imtrt004

1903740
02:31
fix: update model to 360m

imtrt004

198a583
02:19
feat: update base free model

imtrt004

67a030a

Feb 27

05:25
perf: switch default LLM to SmolLM2-1.7B - 40-50% faster tok/s, better instruction following

imtrt004

18dc770
04:57
perf: greedy decoding + dtype fix - 2-3x faster inference on CPU

imtrt004

d16b829
04:43
fix: rewrite loader.py as clean UTF-8 - remove Windows-1252 em-dashes causing SyntaxError

imtrt004

8210d54
04:38
feat: self-hosted Qwen2.5-1.5B-Instruct via transformers — no external API, no compilation

imtrt004

deea70e
04:33
feat: replace llama-cpp-python/Groq with free HF InferenceClient (zero compilation)

imtrt004

98e3f05
01:33
fix: double .gguf extension — skip symlink when path already ends in .gguf; add verbose step logging

imtrt004

dbce995
01:24
feat: LLM readiness tracking — 503 while loading, llm_ready in /health

imtrt004

256f0fc
00:23
fix: symlink blob to .gguf extension so llama.cpp C loader accepts it

imtrt004

915613c
00:08
fix: use hf_hub_download + realpath to avoid snapshot ./path crash

imtrt004

fd0d531

Feb 26

23:50
fix: correct GGUF filename case — Qwen3-4B-Q4_K_M.gguf

imtrt004

1057c73
23:18
Initial backend

imtrt004

b5be2eb