Mirrors
/
hf_text-generation-inference
mirror of https://github.com/huggingface/text-generation-inference.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Impl simple mamba model (#1480) 

This draft PR is a work in progress implementation of the mamba model.
This PR currently loads weights, and produces correct logits after a
single pass.

This PR still needs to correctly integrate this model so it produces
tokens as expected, and apply optimization to avoid all copies during
runtime/unnecessary operations.

#### Helpful resources
[Mamba: Linear-Time Sequence Modeling with Selective State Spaces
(Albert Gu and Tri Dao)](https://arxiv.org/abs/2312.00752)
https://github.com/johnma2006/mamba-minimal

https://github.com/huggingface/candle/blob/main/candle-examples/examples/mamba-minimal/model.rs
https://github.com/huggingface/transformers/pull/28094

Notes: this dev work is currently targeting `state-spaces/mamba-130m`,
so if you want to test please use that model. Additionally when starting
the router the prefill needs to be limited: `cargo run --
--max-batch-prefill-tokens 768 --max-input-length 768`


## Update / Current State

Integration tests have been added and basic functionality such as model
loading is supported.

```bash
cd integration-tests
pytest -vv models/test_fused_kernel_mamba.py
```
- [x] add tests
- [x] load model
- [x] make simple request 
- [ ] resolve warmup issue
- [ ] resolve output issues


fetching models tested during dev
```bash
text-generation-server download-weights state-spaces/mamba-130m
text-generation-server download-weights state-spaces/mamba-1.4b
text-generation-server download-weights state-spaces/mamba-2.8b
```

The server can be run 
```bash
cd server
 MASTER_ADDR=127.0.0.1 MASTER_PORT=5555 python text_generation_server/cli.py serve state-spaces/mamba-2.8b
```

router
```bash
cargo run
```

make a request
```bash
curl -s localhost:3000/generate \
    -X POST \
    -d '{"inputs":"What is Deep Learning?","parameters":{"max_new_tokens":20}}' \
    -H 'Content-Type: application/json' | jq
```

response
```json
{
  "generated_text": "\n\nDeep learning is a machine learning technique that uses a deep neural network to learn from data."
}
```

---------

Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
This commit is contained in:
drbh 2024-02-08 04:19:45 -05:00 committed by GitHub
parent 1734540211
commit bd405e035b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
11 changed files with 1547 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
Dockerfile
View File

@ -154,6 +154,12 @@ COPY server/Makefile-vllm Makefile
# Build specific version of vllm
RUN make build-vllm-cuda
# Build mamba kernels
FROM kernel-builder as mamba-builder
WORKDIR /usr/src
COPY server/Makefile-selective-scan Makefile
RUN make build-all
# Build megablocks
FROM kernel-builder as megablocks-builder
@ -205,6 +211,10 @@ COPY --from=eetq-kernels-builder /usr/src/eetq/build/lib.linux-x86_64-cpython-31
# Copy builds artifacts from vllm builder
COPY --from=vllm-builder /usr/src/vllm/build/lib.linux-x86_64-cpython-310 /opt/conda/lib/python3.10/site-packages
# Copy build artifacts from mamba builder
COPY --from=mamba-builder /usr/src/mamba/build/lib.linux-x86_64-cpython-310/ /opt/conda/lib/python3.10/site-packages
COPY --from=mamba-builder /usr/src/causal-conv1d/build/lib.linux-x86_64-cpython-310/ /opt/conda/lib/python3.10/site-packages
# Install flash-attention dependencies
RUN pip install einops --no-cache-dir

73
integration-tests/models/__snapshots__/test_mamba/test_mamba.json Normal file
View File

@ -0,0 +1,73 @@
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [],
"seed": null,
"tokens": [
{
"id": 187,
"logprob": -0.3552246,
"special": false,
"text": "\n"
},
{
"id": 187,
"logprob": -0.38378906,
"special": false,
"text": "\n"
},
{
"id": 30763,
"logprob": -1.140625,
"special": false,
"text": "Deep"
},
{
"id": 4715,
"logprob": -0.5551758,
"special": false,
"text": " learning"
},
{
"id": 310,
"logprob": -0.59033203,
"special": false,
"text": " is"
},
{
"id": 247,
"logprob": -0.70654297,
"special": false,
"text": " a"
},
{
"id": 747,
"logprob": -2.0410156,
"special": false,
"text": " new"
},
{
"id": 1511,
"logprob": -2.3789062,
"special": false,
"text": " type"
},
{
"id": 273,
"logprob": -0.0026435852,
"special": false,
"text": " of"
},
{
"id": 5145,
"logprob": -1.2841797,
"special": false,
"text": " machine"
}
],
"top_tokens": null
},
"generated_text": "\n\nDeep learning is a new type of machine"
}

99
integration-tests/models/__snapshots__/test_mamba/test_mamba_all_params.json Normal file
View File

@ -0,0 +1,99 @@
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2502,
"logprob": null,
"text": " red"
},
{
"id": 13,
"logprob": -2.5234375,
"text": ","
},
{
"id": 8862,
"logprob": -3.4433594,
"text": " yellow"
},
{
"id": 13,
"logprob": -0.43017578,
"text": ","
},
{
"id": 209,
"logprob": -8.21875,
"text": " "
}
],
"seed": 0,
"tokens": [
{
"id": 187,
"logprob": 0.0,
"special": false,
"text": "\n"
},
{
"id": 395,
"logprob": -0.46411133,
"special": false,
"text": "and"
},
{
"id": 13735,
"logprob": -2.1132812,
"special": false,
"text": " orange"
},
{
"id": 313,
"logprob": -1.2128906,
"special": false,
"text": " ("
},
{
"id": 249,
"logprob": -2.3671875,
"special": false,
"text": "in"
},
{
"id": 253,
"logprob": 0.0,
"special": false,
"text": " the"
},
{
"id": 1340,
"logprob": -1.640625,
"special": false,
"text": " order"
},
{
"id": 597,
"logprob": -0.5488281,
"special": false,
"text": " they"
},
{
"id": 3176,
"logprob": -0.48608398,
"special": false,
"text": " appear"
},
{
"id": 275,
"logprob": 0.0,
"special": false,
"text": " in"
}
],
"top_tokens": null
},
"generated_text": "blue, red, yellow, \nand orange (in the order they appear in"
}

398
integration-tests/models/__snapshots__/test_mamba/test_mamba_load.json Normal file
View File

@ -0,0 +1,398 @@
[
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 1276,
"logprob": null,
"text": "What"
},
{
"id": 310,
"logprob": -0.8125,
"text": " is"
},
{
"id": 18147,
"logprob": -12.828125,
"text": " Deep"
},
{
"id": 20727,
"logprob": -3.0,
"text": " Learning"
},
{
"id": 32,
"logprob": -1.1484375,
"text": "?"
}
],
"seed": null,
"tokens": [
{
"id": 187,
"logprob": -0.3552246,
"special": false,
"text": "\n"
},
{
"id": 187,
"logprob": -0.38378906,
"special": false,
"text": "\n"
},
{
"id": 30763,
"logprob": -1.1279297,
"special": false,
"text": "Deep"
},
{
"id": 4715,
"logprob": -0.5595703,
"special": false,
"text": " learning"
},
{
"id": 310,
"logprob": -0.60253906,
"special": false,
"text": " is"
},
{
"id": 247,
"logprob": -0.7050781,
"special": false,
"text": " a"
},
{
"id": 747,
"logprob": -2.0488281,
"special": false,
"text": " new"
},
{
"id": 1511,
"logprob": -2.3808594,
"special": false,
"text": " type"
},
{
"id": 273,
"logprob": -0.0026416779,
"special": false,
"text": " of"
},
{
"id": 5145,
"logprob": -1.2851562,
"special": false,
"text": " machine"
}
],
"top_tokens": null
},
"generated_text": "\n\nDeep learning is a new type of machine"
},
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 1276,
"logprob": null,
"text": "What"
},
{
"id": 310,
"logprob": -0.78027344,
"text": " is"
},
{
"id": 18147,
"logprob": -12.8203125,
"text": " Deep"
},
{
"id": 20727,
"logprob": -2.9902344,
"text": " Learning"
},
{
"id": 32,
"logprob": -1.1523438,
"text": "?"
}
],
"seed": null,
"tokens": [
{
"id": 187,
"logprob": -0.35351562,
"special": false,
"text": "\n"
},
{
"id": 187,
"logprob": -0.38256836,
"special": false,
"text": "\n"
},
{
"id": 30763,
"logprob": -1.1269531,
"special": false,
"text": "Deep"
},
{
"id": 4715,
"logprob": -0.54541016,
"special": false,
"text": " learning"
},
{
"id": 310,
"logprob": -0.59765625,
"special": false,
"text": " is"
},
{
"id": 247,
"logprob": -0.7001953,
"special": false,
"text": " a"
},
{
"id": 747,
"logprob": -2.0585938,
"special": false,
"text": " new"
},
{
"id": 1511,
"logprob": -2.3789062,
"special": false,
"text": " type"
},
{
"id": 273,
"logprob": -0.0027446747,
"special": false,
"text": " of"
},
{
"id": 5145,
"logprob": -1.2851562,
"special": false,
"text": " machine"
}
],
"top_tokens": null
},
"generated_text": "\n\nDeep learning is a new type of machine"
},
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 1276,
"logprob": null,
"text": "What"
},
{
"id": 310,
"logprob": -0.78027344,
"text": " is"
},
{
"id": 18147,
"logprob": -12.8203125,
"text": " Deep"
},
{
"id": 20727,
"logprob": -2.9902344,
"text": " Learning"
},
{
"id": 32,
"logprob": -1.1523438,
"text": "?"
}
],
"seed": null,
"tokens": [
{
"id": 187,
"logprob": -0.35351562,
"special": false,
"text": "\n"
},
{
"id": 187,
"logprob": -0.38256836,
"special": false,
"text": "\n"
},
{
"id": 30763,
"logprob": -1.1269531,
"special": false,
"text": "Deep"
},
{
"id": 4715,
"logprob": -0.54541016,
"special": false,
"text": " learning"
},
{
"id": 310,
"logprob": -0.59765625,
"special": false,
"text": " is"
},
{
"id": 247,
"logprob": -0.7001953,
"special": false,
"text": " a"
},
{
"id": 747,
"logprob": -2.0585938,
"special": false,
"text": " new"
},
{
"id": 1511,
"logprob": -2.3789062,
"special": false,
"text": " type"
},
{
"id": 273,
"logprob": -0.0027446747,
"special": false,
"text": " of"
},
{
"id": 5145,
"logprob": -1.2851562,
"special": false,
"text": " machine"
}
],
"top_tokens": null
},
"generated_text": "\n\nDeep learning is a new type of machine"
},
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 1276,
"logprob": null,
"text": "What"
},
{
"id": 310,
"logprob": -0.78027344,
"text": " is"
},
{
"id": 18147,
"logprob": -12.8203125,
"text": " Deep"
},
{
"id": 20727,
"logprob": -2.9902344,
"text": " Learning"
},
{
"id": 32,
"logprob": -1.1523438,
"text": "?"
}
],
"seed": null,
"tokens": [
{
"id": 187,
"logprob": -0.35351562,
"special": false,
"text": "\n"
},
{
"id": 187,
"logprob": -0.38256836,
"special": false,
"text": "\n"
},
{
"id": 30763,
"logprob": -1.1269531,
"special": false,
"text": "Deep"
},
{
"id": 4715,
"logprob": -0.54541016,
"special": false,
"text": " learning"
},
{
"id": 310,
"logprob": -0.59765625,
"special": false,
"text": " is"
},
{
"id": 247,
"logprob": -0.7001953,
"special": false,
"text": " a"
},
{
"id": 747,
"logprob": -2.0585938,
"special": false,
"text": " new"
},
{
"id": 1511,
"logprob": -2.3789062,
"special": false,
"text": " type"
},
{
"id": 273,
"logprob": -0.0027446747,
"special": false,
"text": " of"
},
{
"id": 5145,
"logprob": -1.2851562,
"special": false,
"text": " machine"
}
],
"top_tokens": null
},
"generated_text": "\n\nDeep learning is a new type of machine"
}
]

59
integration-tests/models/test_mamba.py Normal file
View File

@ -0,0 +1,59 @@
import pytest
@pytest.fixture(scope="module")
def fused_kernel_mamba_handle(launcher):
with launcher("state-spaces/mamba-130m", num_shard=1) as handle:
yield handle
@pytest.fixture(scope="module")
async def fused_kernel_mamba(fused_kernel_mamba_handle):
await fused_kernel_mamba_handle.health(300)
return fused_kernel_mamba_handle.client
@pytest.mark.asyncio
@pytest.mark.private
async def test_mamba(fused_kernel_mamba, response_snapshot):
response = await fused_kernel_mamba.generate(
"What is Deep Learning?", max_new_tokens=10
)
assert response.details.generated_tokens == 10
assert response.generated_text == "\n\nDeep learning is a new type of machine"
assert response == response_snapshot
@pytest.mark.asyncio
@pytest.mark.private
async def test_mamba_all_params(fused_kernel_mamba, response_snapshot):
response = await fused_kernel_mamba.generate(
"blue, red, yellow, ",
max_new_tokens=10,
repetition_penalty=1.2,
return_full_text=True,
stop_sequences=["test"],
temperature=0.5,
top_p=0.9,
top_k=10,
truncate=5,
typical_p=0.9,
watermark=True,
decoder_input_details=True,
seed=0,
)
assert response.details.generated_tokens == 10
assert response.generated_text == "blue, red, yellow, \nand orange (in the order they appear in"
assert response == response_snapshot
@pytest.mark.asyncio
@pytest.mark.private
async def test_mamba_load(fused_kernel_mamba, generate_load, response_snapshot):
responses = await generate_load(fused_kernel_mamba, "What is Deep Learning?", max_new_tokens=10, n=4)
assert len(responses) == 4
assert all([r.generated_text == responses[0].generated_text for r in responses])
assert responses[0].generated_text == "\n\nDeep learning is a new type of machine"
assert responses == response_snapshot

1
server/.gitignore vendored
View File

@ -161,3 +161,4 @@ flash-attention-v2/
vllm/
llm-awq/
eetq/
mamba/

1
server/Makefile
View File

@ -3,6 +3,7 @@ include Makefile-flash-att-v2
include Makefile-vllm
include Makefile-awq
include Makefile-eetq
include Makefile-selective-scan
unit-tests:
pytest -s -vv -m "not private" tests

28
server/Makefile-selective-scan Normal file
View File

@ -0,0 +1,28 @@
selective_scan_commit := 2a3704fd47ba817b415627b06fd796b971fdc137
causal-conv1d:
rm -rf causal-conv1d
git clone https://github.com/Dao-AILab/causal-conv1d.git
build-causal-conv1d: causal-conv1d
cd causal-conv1d/ && git checkout v1.1.1 # known latest working version tag
cd causal-conv1d/ && CAUSAL_CONV1D_FORCE_BUILD=TRUE python setup.py build
install-causal-conv1d: build-causal-conv1d
pip uninstall causal-conv1d -y || true
cd causal-conv1d/ && pip install .
# selective-scan dependends on causal-conv1d
selective-scan:
rm -rf mamba
git clone https://github.com/state-spaces/mamba.git mamba
build-selective-scan: selective-scan
cd mamba/ && git fetch && git checkout $(selective_scan_commit)
cd mamba && python setup.py build
install-selective-scan: install-causal-conv1d build-selective-scan
pip uninstall selective-scan-cuda -y || true
cd mamba && pip install .
build-all: build-causal-conv1d build-selective-scan

29
server/text_generation_server/models/__init__.py
View File

@ -76,6 +76,15 @@ if FLASH_ATTENTION:
__all__.append(FlashMixtral)
__all__.append(FlashPhi)
MAMBA_AVAILABLE = True
try:
from text_generation_server.models.mamba import Mamba
except ImportError as e:
logger.warning(f"Could not import Mamba: {e}")
MAMBA_AVAILABLE = False
if MAMBA_AVAILABLE:
__all__.append(Mamba)
def get_model(
model_id: str,
@ -164,7 +173,25 @@ def get_model(
if speculate > 0:
logger.info(f"Using speculation {method} with {speculate} input ids.")
model_type = config_dict["model_type"]
model_type = config_dict.get("model_type", None)
if model_type is None:
# TODO: fix how we determine model type for Mamba
if "ssm_cfg" in config_dict:
# *only happens in Mamba case
model_type = "ssm"
else:
raise RuntimeError(
f"Could not determine model type for {model_id} revision {revision}"
)
if model_type == "ssm":
return Mamba(
model_id,
revision,
quantize=quantize,
dtype=dtype,
trust_remote_code=trust_remote_code,
)
if model_type == "gpt_bigcode":
if FLASH_ATTENTION:

194
server/text_generation_server/models/custom_modeling/mamba_modeling.py Normal file
View File

@ -0,0 +1,194 @@
import torch
import torch.distributed
from mamba_ssm.ops.triton.selective_state_update import selective_state_update
from mamba_ssm.ops.selective_scan_interface import selective_scan_fn
from mamba_ssm.utils.generation import InferenceParams
from torch import nn
from typing import Optional, Tuple, Any
from transformers.configuration_utils import PretrainedConfig
import torch.nn.functional as F
from text_generation_server.utils.layers import (
TensorParallelEmbedding,
FastRMSNorm,
FastLinear,
)
from einops import rearrange
from causal_conv1d import causal_conv1d_fn, causal_conv1d_update
import math
class MambaConfig(PretrainedConfig):
def __init__(
self,
vocab_size=50280,
d_model=768,
d_state=16,
n_layer=32,
layer_norm_epsilon=1e-5,
tie_word_embeddings=False,
pad_token_id=0,
bos_token_id=1,
eos_token_id=2,
expand=2,
dt_rank="auto",
**kwargs,
):
self.vocab_size = vocab_size
self.n_layer = n_layer
self.layer_norm_epsilon = layer_norm_epsilon
self.d_model = d_model
self.d_inner = d_model * 2
self.d_conv = 4
self.d_state = d_state
self.expand = expand
self.dt_rank = math.ceil(self.d_model / 16) if dt_rank == "auto" else dt_rank
super().__init__(
pad_token_id=pad_token_id,
bos_token_id=bos_token_id,
eos_token_id=eos_token_id,
tie_word_embeddings=tie_word_embeddings,
**kwargs,
)
class MambaBlock(nn.Module):
def __init__(self, prefix, config, weights):
super().__init__()
self.layer_idx = int(prefix.split(".")[2])
self.in_proj = FastLinear.load(config, f"{prefix}.in_proj", weights, bias=False)
self.x_proj = FastLinear.load(config, f"{prefix}.x_proj", weights, bias=False)
self.dt_proj = FastLinear.load(config, f"{prefix}.dt_proj", weights, bias=True)
self.dt_proj_no_bias = FastLinear.load(config, f"{prefix}.dt_proj", weights, bias=False)
self.out_proj = FastLinear.load(config, f"{prefix}.out_proj", weights, bias=False)
self.conv1d = FastLinear.load(config, f"{prefix}.conv1d", weights, bias=True)
self.negA = -torch.exp(weights.get_tensor(f"{prefix}.A_log").float())
self.D = weights.get_tensor(f"{prefix}.D")
self.activation = "silu"
self.dt_rank = config.dt_rank
self.d_state = config.d_state
self.d_conv = config.d_conv
self.act = nn.SiLU()
# inference_params
def forward(self, hidden_states: torch.Tensor, inference_params=None):
_, seqlen, _ = hidden_states.shape
conv_state, ssm_state = inference_params.key_value_memory_dict[self.layer_idx]
if inference_params.seqlen_offset > 0:
out, conv_state, ssm_state = self.step(hidden_states, conv_state, ssm_state)
return out, conv_state, ssm_state
projected_states = self.in_proj(hidden_states).transpose(1,2)
x, z = projected_states.chunk(2, dim=1)
conv_state = F.pad(x, (self.d_conv - seqlen, 0))
x = causal_conv1d_fn(
x=x,
weight=self.conv1d.weight.view(self.conv1d.weight.size(0), self.conv1d.weight.size(2)),
bias=self.conv1d.bias,
activation=self.activation,
)
# We're careful here about the layout, to avoid extra transposes.
# We want dt to have d as the slowest moving dimension
# and L as the fastest moving dimension, since those are what the ssm_scan kernel expects.
x_dbl = self.x_proj(rearrange(x, "b d l -> (b l) d")) # (bl d)
dt, B, C = torch.split(x_dbl, [self.dt_rank, self.d_state, self.d_state], dim=-1)
dt = self.dt_proj.weight @ dt.t()
dt = rearrange(dt, "d (b l) -> b d l", l=seqlen)
B = rearrange(B, "(b l) dstate -> b dstate l", l=seqlen).contiguous()
C = rearrange(C, "(b l) dstate -> b dstate l", l=seqlen).contiguous()
y, last_state = selective_scan_fn(
x,
dt,
self.negA,
B,
C,
self.D.float(),
z=z,
delta_bias=self.dt_proj.bias.float(),
delta_softplus=True,
return_last_state=True,
)
y = rearrange(y, "b d l -> b l d")
attn_outputs = self.out_proj(y)
return attn_outputs, conv_state, last_state
def step(self, hidden_states, conv_state, ssm_state):
_xz = self.in_proj(hidden_states)
_x, _z = _xz.chunk(2, dim=-1) # (B D)
conv_state_new = torch.cat([conv_state, _x.transpose(1,2)], dim=-1)
conv_out = causal_conv1d_fn(
x=conv_state_new,
weight=self.conv1d.weight.view(self.conv1d.weight.size(0), self.conv1d.weight.size(2)),
bias=self.conv1d.bias,
activation=self.activation
)
conv_state = conv_state_new[:, :, 1:]
bsz, seqlen, dim = hidden_states.shape
output_tensor = torch.zeros(
(bsz, seqlen, dim),
device=hidden_states.device,
dtype=hidden_states.dtype
)
for i in range(0, bsz):
x = conv_out[i:i+1,:,-1]
z = _z[i:i+1, -1, :]
x_db = self.x_proj(x)
dt, B, C = torch.split(x_db, [self.dt_rank, self.d_state, self.d_state], dim=-1)
dt = F.linear(dt, self.dt_proj.weight)
y = selective_state_update(
ssm_state[i:i+1,:,:], x, dt, self.negA, B, C, self.D, z=z, dt_bias=self.dt_proj.bias, dt_softplus=True
)
out = self.out_proj(y)
output_tensor[i] = out
return output_tensor, conv_state, ssm_state
class ResidualBlock(nn.Module):
def __init__(self, layer_id, config, weights):
super().__init__()
self.mamba_block = MambaBlock(prefix=f"{layer_id}.mixer", config=config, weights=weights)
self.layer_norm = FastRMSNorm.load(prefix=f"{layer_id}.norm", weights=weights, eps=config.layer_norm_epsilon)
def forward(
self,
hidden_states: torch.Tensor,
residual: Optional[torch.Tensor] = None,
inference_params: Optional[Any] = None,
):
residual = (hidden_states + residual) if residual is not None else hidden_states
shape = residual.shape
hidden_states, _ = self.layer_norm(residual.view(-1, shape[-1]))
hidden_states, conv_state, last_ssm_state = self.mamba_block(hidden_states.view(*shape), inference_params)
return hidden_states, residual, conv_state, last_ssm_state
class MambaModel(nn.Module):
def __init__(self, config, weights):
super().__init__()
prefix = "backbone"
self.embed_tokens = TensorParallelEmbedding(f"{prefix}.embedding", weights)
self.blocks = nn.ModuleList(
[ResidualBlock(f"{prefix}.layers.{i}", config, weights) for i in range(config.n_layer)]
)
self.norm_f = FastRMSNorm.load(f"{prefix}.norm_f", weights, eps=config.layer_norm_epsilon)
self.lm_head = FastLinear.load(config, f"{prefix}.embedding", weights, bias=False)
self.config = config
def forward(self, input_ids: torch.Tensor, inference_params=None, residual=None) -> Tuple[torch.Tensor, torch.Tensor, InferenceParams]:
hidden_states = self.embed_tokens(input_ids)
for block in self.blocks:
hidden_states, residual, conv_state, ssm_state = block(hidden_states, residual, inference_params)
inference_params.key_value_memory_dict[block.mamba_block.layer_idx] = (conv_state, ssm_state)
hidden_states = hidden_states + residual if residual is not None else hidden_states
hidden_states, _ = self.norm_f(hidden_states.view(-1, hidden_states.size(-1)))
hidden_states = hidden_states.view(residual.shape)
logits = self.lm_head(hidden_states)
# update the offset for the next inference using these params
inference_params.seqlen_offset += input_ids.size(1)
return logits, input_ids, inference_params

656
server/text_generation_server/models/mamba.py Normal file
View File

@ -0,0 +1,656 @@
import torch
import torch.distributed
from transformers import AutoTokenizer, PreTrainedTokenizerBase
from typing import Optional
from text_generation_server.models.custom_modeling.mamba_modeling import (
MambaConfig,
)
from text_generation_server.pb import generate_pb2
from text_generation_server.utils import (
initialize_torch_distributed,
weight_files,
Weights,
)
import time
from text_generation_server.models.custom_modeling.mamba_modeling import MambaModel
from text_generation_server.models import Model
from typing import Any, List, Optional, Tuple, Type, Dict
from text_generation_server.models.types import (
Batch,
Tokens,
Generation,
GeneratedText,
)
from text_generation_server.utils.tokens import batch_top_tokens, Sampling
from dataclasses import dataclass
from text_generation_server.utils import NextTokenChooser, StoppingCriteria, Sampling
from mamba_ssm.utils.generation import InferenceParams
@dataclass
class MambaBatch(Batch):
batch_id: int
requests: List[generate_pb2.Request]
requests_idx_mapping: Dict[int, int]
# Decoder values
input_ids: torch.Tensor
# All tokens
all_input_ids: List[torch.Tensor]
# Lengths of all generations present in the batch
input_lengths: List[int]
prefix_offsets: List[int]
read_offsets: List[int]
# Generation helpers
next_token_choosers: List[NextTokenChooser]
stopping_criterias: List[StoppingCriteria]
top_n_tokens: List[int]
top_n_tokens_tensor: torch.Tensor
# Metadata used for padding
max_input_length: int
padding_right_offset: int
# Maximum number of tokens this batch will grow to
max_tokens: int
# Past metadata
keys_head_dim_last: bool = True
# Inference params
inference_params: Optional[Dict[str, Any]] = None
def to_pb(self) -> generate_pb2.CachedBatch:
return generate_pb2.CachedBatch(
id=self.batch_id,
request_ids=[r.id for r in self.requests],
size=len(self),
max_tokens=self.max_tokens,
)
@classmethod
def from_pb(
cls,
pb: generate_pb2.Batch,
tokenizer: PreTrainedTokenizerBase,
dtype: torch.dtype,
device: torch.device,
) -> "MambaBatch":
inputs = []
next_token_choosers = []
stopping_criterias = []
top_n_tokens = []
prefix_offsets = []
read_offsets = []
requests_idx_mapping = {}
# Parse batch
max_truncation = 0
padding_right_offset = 0
max_decode_tokens = 0
for i, r in enumerate(pb.requests):
requests_idx_mapping[r.id] = i
inputs.append(r.inputs)
next_token_choosers.append(NextTokenChooser.from_pb(r.parameters, device))
stopping_criteria = StoppingCriteria.from_pb(
r.stopping_parameters, tokenizer
)
stopping_criterias.append(stopping_criteria)
top_n_tokens.append(r.top_n_tokens)
max_truncation = max(max_truncation, r.truncate)
max_decode_tokens += stopping_criteria.max_new_tokens
padding_right_offset = max(
padding_right_offset, stopping_criteria.max_new_tokens
)
tokenized_inputs = tokenizer(
inputs,
return_tensors="pt",
padding=True,
return_token_type_ids=False,
truncation=True,
max_length=max_truncation,
).to(device)
for _ in pb.requests:
input_len = tokenized_inputs["input_ids"].shape[1]
prefix_offsets.append(input_len - 5)
read_offsets.append(input_len)
input_lengths = tokenized_inputs["attention_mask"].sum(1)
max_input_length = input_lengths.max()
input_ids = tokenized_inputs["input_ids"]
all_input_ids = tokenized_inputs["input_ids"].T.split(1, dim=1)
top_n_tokens_tensor = torch.tensor(
top_n_tokens, device=device, dtype=torch.int64
)
max_tokens = len(inputs) * (max_input_length + max_decode_tokens)
return cls(
batch_id=pb.id,
requests=pb.requests,
requests_idx_mapping=requests_idx_mapping,
input_ids=input_ids,
# past_input_ids=None,
all_input_ids=list(all_input_ids),
input_lengths=input_lengths.tolist(),
prefix_offsets=prefix_offsets,
read_offsets=read_offsets,
next_token_choosers=next_token_choosers,
stopping_criterias=stopping_criterias,
top_n_tokens=top_n_tokens,
top_n_tokens_tensor=top_n_tokens_tensor,
max_input_length=max_input_length.item(),
padding_right_offset=padding_right_offset,
max_tokens=max_tokens,
)
def filter(self, request_ids: List[int]) -> Optional["MambaBatch"]:
if len(request_ids) == 0:
raise ValueError("Batch must have at least one request")
if len(request_ids) == len(self):
return self
keep_indices = []
# New values after filtering
requests_idx_mapping = {}
requests = []
input_lengths = []
prefix_offsets = []
read_offsets = []
all_input_ids = []
max_input_length = 0
next_token_choosers = []
stopping_criterias = []
top_n_tokens = []
total_remaining_decode_tokens = 0
new_padding_right_offset = 0
indices = []
for i, request_id in enumerate(request_ids):
idx = self.requests_idx_mapping[request_id]
requests_idx_mapping[request_id] = i
keep_indices.append(idx)
requests.append(self.requests[idx])
prefix_offsets.append(self.prefix_offsets[idx])
read_offsets.append(self.read_offsets[idx])
all_input_ids.append(self.all_input_ids[idx])
request_input_length = self.input_lengths[idx]
input_lengths.append(request_input_length)
max_input_length = max(max_input_length, request_input_length)
indices.append(idx)
next_token_choosers.append(self.next_token_choosers[idx])
stopping_criteria = self.stopping_criterias[idx]
stopping_criterias.append(stopping_criteria)
top_n_tokens.append(self.top_n_tokens[idx])
remaining_decode_tokens = (
stopping_criteria.max_new_tokens - stopping_criteria.current_tokens
)
total_remaining_decode_tokens += remaining_decode_tokens
new_padding_right_offset = max(
new_padding_right_offset, remaining_decode_tokens
)
# Apply indices to input_ids, attention mask, past key values and other items that need to be cached
input_ids = self.input_ids[keep_indices]
top_n_tokens_tensor = self.top_n_tokens_tensor[keep_indices]
max_tokens = len(request_ids) * max_input_length + total_remaining_decode_tokens
self.requests = requests
self.requests_idx_mapping = requests_idx_mapping
self.input_ids = input_ids
self.all_input_ids = all_input_ids
self.input_lengths = input_lengths
self.prefix_offsets = prefix_offsets
self.read_offsets = read_offsets
self.next_token_choosers = next_token_choosers
self.stopping_criterias = stopping_criterias
self.top_n_tokens = top_n_tokens
self.top_n_tokens_tensor = top_n_tokens_tensor
self.max_input_length = max_input_length
self.padding_right_offset = new_padding_right_offset
self.max_tokens = max_tokens
# TODO
# Kept it simple by just updating the state, maybe updating the other CPU values is necessary.
key_value_memory_dict = {}
for i, (conv_state, ssm_state) in self.inference_params.key_value_memory_dict.items():
key_value_memory_dict[i] = (conv_state[indices], ssm_state[indices])
self.inference_params.key_value_memory_dict = key_value_memory_dict
return self
@classmethod
def concatenate(cls, batches: List["MambaBatch"]) -> "MambaBatch":
# Used for padding
total_batch_size = 0
max_input_length = 0
padding_right_offset = 0
for batch in batches:
total_batch_size += len(batch)
max_input_length = max(max_input_length, batch.max_input_length)
padding_right_offset = max(padding_right_offset, batch.padding_right_offset)
# Batch attributes
requests = []
requests_idx_mapping = {}
input_lengths = []
prefix_offsets = []
read_offsets = []
all_input_ids = []
next_token_choosers = []
stopping_criterias = []
top_n_tokens = []
max_tokens = 0
max_seqlen = 0
batch_size = 0
seqlen_offset = 0
# Batch tensors
input_ids = None
top_n_tokens_tensor = None
# Used for slicing correctly inside the tensors
# Equivalent to a cumsum on batch sizes
start_index = 0
for i, batch in enumerate(batches):
requests.extend(batch.requests)
input_lengths.extend(batch.input_lengths)
prefix_offsets.extend(batch.prefix_offsets)
read_offsets.extend(batch.read_offsets)
all_input_ids.extend(batch.all_input_ids)
next_token_choosers.extend(batch.next_token_choosers)
stopping_criterias.extend(batch.stopping_criterias)
top_n_tokens.extend(batch.top_n_tokens)
if i == 0:
requests_idx_mapping = batch.requests_idx_mapping
else:
# We need to offset the mapping for each batch by the cumulative batch size
for k, v in batch.requests_idx_mapping.items():
requests_idx_mapping[k] = v + start_index
# Slicing end index for this batch
end_index = start_index + len(batch)
# Create empty tensor
# input_ids is always of shape [batch_size, 1]
# We do not need to pad it
if input_ids is None:
input_ids = batch.input_ids.new_empty((total_batch_size, 1))
# Copy to correct indices
input_ids[start_index:end_index] = batch.input_ids
if top_n_tokens_tensor is None:
top_n_tokens_tensor = batches[0].top_n_tokens_tensor.new_zeros(
total_batch_size,
)
top_n_tokens_tensor[start_index:end_index] = batch.top_n_tokens_tensor
# Add eventual padding tokens that were added while concatenating
max_tokens += batch.max_tokens + (
max_input_length - batch.max_input_length
) * len(batch)
max_seqlen = max(max_seqlen, batch.inference_params.max_seqlen)
seqlen_offset = max(seqlen_offset, batch.inference_params.seqlen_offset)
batch_size += batch.inference_params.max_batch_size
start_index = end_index
(_, d_model, d_conv) = batches[0].inference_params.key_value_memory_dict[0][0].shape
(_, _, d_state) = batches[0].inference_params.key_value_memory_dict[0][1].shape
n_blocks = len(batches[0].inference_params.key_value_memory_dict)
dtype = batches[0].inference_params.key_value_memory_dict[0][0].dtype
device = batches[0].inference_params.key_value_memory_dict[0][0].device
key_value_memory_dict = {}
for i in range(n_blocks):
conv_state = torch.zeros(
batch_size,
d_model,
d_conv,
device=device,
dtype=dtype,
)
ssm_state = torch.zeros(
batch_size,
d_model,
d_state,
device=device,
dtype=dtype,
)
key_value_memory_dict[i] = (conv_state, ssm_state)
lengths_per_sample = torch.zeros(batch_size, dtype=torch.int32, device=device)
inference_params = InferenceParams(
max_seqlen=max_seqlen,
max_batch_size=batch_size,
seqlen_offset=seqlen_offset,
key_value_memory_dict=key_value_memory_dict,
lengths_per_sample=lengths_per_sample,
)
current_batch = 0
for batch in batches:
for i in range(n_blocks):
conv_state, ssm_state = batch.inference_params.key_value_memory_dict[i]
batch_size = batch.inference_params.max_batch_size
inference_params.key_value_memory_dict[i][0][current_batch:current_batch + batch_size] = conv_state
inference_params.key_value_memory_dict[i][1][current_batch:current_batch + batch_size] = ssm_state
inference_params.lengths_per_sample[current_batch: current_batch + batch_size] = batch.inference_params.lengths_per_sample
current_batch += batch_size
return cls(
batch_id=batches[0].batch_id,
requests=requests,
requests_idx_mapping=requests_idx_mapping,
input_ids=input_ids,
all_input_ids=all_input_ids,
input_lengths=input_lengths,
prefix_offsets=prefix_offsets,
read_offsets=read_offsets,
next_token_choosers=next_token_choosers,
stopping_criterias=stopping_criterias,
top_n_tokens=top_n_tokens,
top_n_tokens_tensor=top_n_tokens_tensor,
max_input_length=max_input_length,
padding_right_offset=padding_right_offset,
keys_head_dim_last=batches[0].keys_head_dim_last,
max_tokens=max_tokens,
inference_params=inference_params
)
def __len__(self):
return len(self.requests)
class Mamba(Model):
def __init__(
self,
model_id: str,
revision: Optional[str] = None,
quantize: Optional[str] = None,
dtype: Optional[torch.dtype] = None,
trust_remote_code: bool = False,
):
self.process_group, _rank, _world_size = initialize_torch_distributed()
if torch.cuda.is_available():
device = torch.device("cuda")
dtype = torch.float16 if dtype is None else dtype
else:
if quantize:
raise ValueError("quantization is not available on CPU")
device = torch.device("cpu")
dtype = torch.float32 if dtype is None else dtype
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/gpt-neox-20b",
revision=revision,
padding_side="left",
truncation_side="left",
trust_remote_code=trust_remote_code,
)
config = MambaConfig.from_pretrained(
model_id, revision=revision, trust_remote_code=trust_remote_code
)
tokenizer.bos_token_id = config.bos_token_id
tokenizer.eos_token_id = config.eos_token_id
tokenizer.pad_token = tokenizer.eos_token
config.quantize = quantize
torch.distributed.barrier(group=self.process_group)
filenames = weight_files(model_id, revision=revision, extension=".safetensors")
weights = Weights(filenames, device, dtype, process_group=self.process_group)
model = MambaModel(config, weights)
torch.distributed.barrier(group=self.process_group)
super(Mamba, self).__init__(
model=model,
tokenizer=tokenizer,
requires_padding=True,
dtype=dtype,
device=device,
)
@property
def batch_type(self) -> Type[MambaBatch]:
return MambaBatch
def warmup(self, batch) -> Optional[int]:
# TODO: implement warmup for Mamba if needed
return None
def forward(
self,
input_ids: torch.Tensor,
past: Optional[List[torch.Tensor]] = None,
) -> Tuple[torch.Tensor, torch.Tensor]:
return self.model(
input_ids,
past=past,
)
def generate_token(self, batch) -> Tuple[List[Any], Optional[Any], Tuple[int, int]]:
start = time.time_ns()
input_ids = batch.input_ids # batch.past_input_ids if batch.past_input_ids is not None else batch.input_ids
batch_size = input_ids.shape[0]
max_seqlen = input_ids.shape[1]
dtype = input_ids.dtype
# Inference params
seqlen_og = 0
inf_cache = {}
lengths_per_sample = torch.ones(batch_size, dtype=torch.int32, device=input_ids.device) * max_seqlen
if batch.inference_params is None:
inference_params = InferenceParams(
max_seqlen=max_seqlen,
max_batch_size=batch_size,
seqlen_offset=seqlen_og,
key_value_memory_dict=inf_cache,
lengths_per_sample=lengths_per_sample,
)
# Allocate inference cache
for res_block in self.model.blocks:
block = res_block.mamba_block
conv_state = torch.zeros(
batch_size,
self.model.config.d_model * self.model.config.expand,
self.model.config.d_conv,
device=block.conv1d.weight.device,
dtype=block.conv1d.weight.dtype,
)
ssm_state = torch.zeros(
batch_size,
self.model.config.d_model * self.model.config.expand,
self.model.config.d_state,
device=block.dt_proj.weight.device,
dtype=block.dt_proj.weight.dtype,
)
inference_params.key_value_memory_dict[block.layer_idx] = (conv_state, ssm_state)
batch.inference_params = inference_params
# Forward pass
logits, past_input_ids, new_inference_params = self.model(input_ids, batch.inference_params)
batch.inference_params = new_inference_params
# Results
generations: List[Generation] = []
stopped = True
# Speculation is not active for causal
accepted_ids = torch.ones_like(batch.input_ids)[:, 0]
batch_top_token_ids, batch_top_token_logprobs = batch_top_tokens(
batch.top_n_tokens,
batch.top_n_tokens_tensor,
torch.log_softmax(logits[:, -1], -1),
accepted_ids,
)
start_decode = time.time_ns()
# Zipped iterator
iterator = zip(
batch.requests,
batch.input_lengths,
batch.prefix_offsets,
batch.read_offsets,
logits,
batch.next_token_choosers,
batch.stopping_criterias,
batch.all_input_ids,
batch.top_n_tokens,
batch_top_token_ids,
batch_top_token_logprobs,
)
# For each member of the batch
for i, (
request,
input_length,
prefix_offset,
read_offset,
logits,
next_token_chooser,
stopping_criteria,
all_input_ids,
top_n_tokens,
top_token_ids,
top_token_logprobs,
) in enumerate(iterator):
# Select next token
next_token_id, logprobs = next_token_chooser(
all_input_ids.view(1, -1), logits[-1:, :]
)
# Append next token to all tokens
all_input_ids = torch.cat([all_input_ids, next_token_id])
new_input_length = input_length + 1
# Generated token
next_token_logprob = logprobs[-1, next_token_id]
next_token_id_squeezed = next_token_id.squeeze()
next_token_text, prefix_offset, read_offset = self.decode_token(
all_input_ids[:, 0], prefix_offset, read_offset
)
# Evaluate stopping criteria
stop, reason = stopping_criteria(
next_token_id_squeezed,
next_token_text,
)
if not stop:
stopped = False
# Shard generations
# All generations will be appended in the rust sharded client
if i % self.world_size == self.rank:
if stop:
# Decode generated tokens
output_text, _, _ = self.decode_token(
all_input_ids[:, 0],
prefix_offset=len(all_input_ids)
- stopping_criteria.current_tokens
- 1,
read_offset=len(all_input_ids) - stopping_criteria.current_tokens,
skip_special_tokens=True,
)
# Get seed
if isinstance(next_token_chooser.choice, Sampling):
seed = next_token_chooser.choice.seed
else:
seed = None
generated_text = GeneratedText(
output_text, stopping_criteria.current_tokens, reason, seed
)
else:
generated_text = None
if stopping_criteria.current_tokens == 1 and request.prefill_logprobs:
# Remove generated token to only have prefill and add nan for first prompt token
prefill_logprobs = [float("nan")] + torch.log_softmax(
logits, -1
).gather(1, all_input_ids[1:]).squeeze(1)[
-new_input_length:-1
].tolist()
prefill_token_ids = all_input_ids[-new_input_length:-1]
prefill_texts = self.tokenizer.batch_decode(
prefill_token_ids,
clean_up_tokenization_spaces=False,
skip_special_tokens=False,
)
prefill_tokens = Tokens(
prefill_token_ids,
prefill_logprobs,
prefill_texts,
is_special=[],
)
else:
prefill_tokens = None
if top_n_tokens > 0:
toptoken_texts = self.tokenizer.batch_decode(
top_token_ids,
clean_up_tokenization_spaces=False,
skip_special_tokens=False,
)
special_toptokens = [
token_id in self.all_special_ids for token_id in top_token_ids
]
top_tokens = Tokens(
top_token_ids,
top_token_logprobs,
toptoken_texts,
special_toptokens,
)
else:
top_tokens = None
generation = Generation(
request.id,
prefill_tokens,
Tokens(
[next_token_id_squeezed],
[next_token_logprob],
[next_token_text],
[next_token_id_squeezed.item() in self.all_special_ids],
),
generated_text,
top_tokens,
)
generations.append(generation)
# Update values
batch.input_ids[i, 0] = next_token_id
batch.all_input_ids[i] = all_input_ids
batch.input_lengths[i] = new_input_length
batch.prefix_offsets[i] = prefix_offset
batch.read_offsets[i] = read_offset
batch.max_input_length = max(batch.max_input_length, new_input_length)
# We finished all generations in the batch; there is no next batch
if stopped:
forward_ns = start_decode - start
decode_ns = time.time_ns() - start_decode
return generations, None, (forward_ns, decode_ns)
# Slice unused values from prefill
batch.input_ids = batch.input_ids[:, :1]
forward_ns = start_decode - start
decode_ns = time.time_ns() - start_decode
return generations, batch, (forward_ns, decode_ns)