Comments (1)
The reason behind this difference in behaviour is because phi3 small is not natively supported whereas mini and medium are. Phi3 small therefore uses the AutoModel implementation that requires a lot more memory and is not ISO in features to native models (no flash/paged attention, padding...).
Phi3 small is really different from the other flavours, from what I could find so far:
- Uses block sparse attention with dense attention every n layers.
- Does not use custom softmax scaling (not sqrt(head_size)).
- Does not use regular gating in the MLP layers, uses the gegelu activation instead, which does gating internally using linear gating. However a big difference is that the gate and non-gate parameters are interleaved (a_gelu, a_linear = input[..., ::2], input[..., 1::2]).
- Residual + layernorm (not RMS) is arranged differently than in Llama.
- Column-packed QKV with GQA
We are considering adding native support for it but have not done it yet.
On a sidenote, phi3 mini is affected by #2055, solved in #2060.
from text-generation-inference.
Related Issues (20)
- Phi-3-mini-128k crashes on simple query HOT 2
- Add Environment Variable for OTLP Service Name HOT 2
- ImportError: libcuda.so.1: cannot open shared object file: No such file or directory HOT 10
- Tree-attention for medusa HOT 2
- get stucked when run text-generation-benchmark on AMD gpu HOT 2
- Unable to load Qwen2-72B-Instruct-exl2 model HOT 2
- `mistralai/Mixtral-8x22B-Instruct-v0.1`: Getting `RuntimeError: 'ptxas' failed with error code 127` while warming up on 8 GPUs HOT 2
- `mistralai/Mixtral-8x22B-Instruct-v0.1`: Successful warmup, crashes on inference HOT 1
- Long install report HOT 1
- P40 with USE_FLASH_ATTENTION=False HOT 2
- Sparse Marlin HOT 3
- protobuf version not compatible HOT 1
- Qwen/Qwen2-72B-Instruct-AWQ gibberish output in 2.0.4 HOT 3
- Will Support Disaggregating Prefill and Decoding? HOT 1
- `top_p` messes up `top_logprobs` HOT 2
- template_error in /chat/completions HOT 2
- how to launch a service using downloaded model weights? HOT 2
- TGI keeps crashing with 'device-side assert triggered' HOT 1
- AttributeError: 'MixtralLayer' object has no attribute 'mlp' HOT 1
- Error loading Qwen2-72B-Instruct with EETQ
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from text-generation-inference.