Cannot load `replit-code-v1-3b`

[matthiasgeihs]

I am using ggml@https://github.com/ggerganov/ggml/commit/f52d2a05cf8327baf6c0d49e7b231953179e03d3 and am having trouble loading replit-code-v1-3b.

I used replit/convert-h5-to-ggml.py to convert to ggml f16. Also created a q4_0 quantized version using replit-quantize.

However, when trying to load either of the models with replit (e.g., ./bin/replit -m replit-code-v1-3b.f16.bin -p "def hello_world():") I get:

replit_model_load: unknown tensor 'transformer.blocks.0.norm_1.weight' in model file
main: failed to load model from '/Users/matthias/.cache/ggml/replit-code-v1-3b.f16.bin'

Any ideas?

[apcameron]

You may need this PR https://github.com/ggerganov/ggml/pull/218

[matthiasgeihs]

Works now on https://github.com/ggerganov/ggml/commit/0a63fc0f6cb1915d1fa5c62c8f0f018d072253c9! 👍