matthayes commited on
Commit
332e267
1 Parent(s): dba64bb

Update README.md

Browse files

Change model name

Files changed (1) hide show
  1. README.md +17 -17
README.md CHANGED
@@ -5,20 +5,20 @@ language:
5
  library_name: transformers
6
  inference: false
7
  ---
8
- # dolly-v2-6-9b Model Card
9
  ## Summary
10
 
11
- Databricks’ `dolly-v2-6-9b`, an instruction-following large language model trained on the Databricks machine learning platform
12
  that is licensed for commercial use. Based on `pythia-6.9b`, Dolly is trained on ~15k instruction/response fine tuning records
13
  [`databricks-dolly-15k`](https://github.com/databrickslabs/dolly/tree/master/data) generated
14
  by Databricks employees in capability domains from the InstructGPT paper, including brainstorming, classification, closed QA, generation,
15
- information extraction, open QA and summarization. `dolly-v2-6-9b` is not a state-of-the-art model, but does exhibit surprisingly
16
  high quality instruction following behavior not characteristic of the foundation model on which it is based.
17
 
18
  **Owner**: Databricks, Inc.
19
 
20
  ## Model Overview
21
- `dolly-v2-6-9b` is a 6.9 billion parameter causal language model created by [Databricks](https://databricks.com/) that is derived from
22
  [EleutherAI’s](https://www.eleuther.ai/) [Pythia-6.9b](https://huggingface.co/EleutherAI/pythia-6.9b) and fine-tuned
23
  on a [~15K record instruction corpus](https://github.com/databrickslabs/dolly/tree/master/data) generated by Databricks employees and released under a permissive license (CC-BY-SA)
24
 
@@ -32,7 +32,7 @@ In a Databricks notebook you could run:
32
  ```
33
 
34
  The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline`
35
- found in the model repo [here](https://huggingface.co/databricks/dolly-v2-6-9b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required.
36
  Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality.
37
  It is also fine to remove it if there is sufficient memory.
38
 
@@ -40,7 +40,7 @@ It is also fine to remove it if there is sufficient memory.
40
  import torch
41
  from transformers import pipeline
42
 
43
- generate_text = pipeline(model="databricks/dolly-v2-6-9b", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto")
44
  ```
45
 
46
  You can then use the pipeline to answer instructions:
@@ -49,15 +49,15 @@ You can then use the pipeline to answer instructions:
49
  generate_text("Explain to me the difference between nuclear fission and fusion.")
50
  ```
51
 
52
- Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-6-9b/blob/main/instruct_pipeline.py),
53
  store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer:
54
 
55
  ```
56
  from instruct_pipeline import InstructionTextGenerationPipeline
57
  from transformers import AutoModelForCausalLM, AutoTokenizer
58
 
59
- tokenizer = AutoTokenizer.from_pretrained("databricks/dolly-v2-6-9b", padding_side="left")
60
- model = AutoModelForCausalLM.from_pretrained("databricks/dolly-v2-6-9b", device_map="auto")
61
 
62
  generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer)
63
  ```
@@ -66,23 +66,23 @@ generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokeniz
66
  ## Known Limitations
67
 
68
  ### Performance Limitations
69
- **`dolly-v2-6-9b` is not a state-of-the-art generative language model** and, though quantitative benchmarking is ongoing, is not designed to perform
70
  competitively with more modern model architectures or models subject to larger pretraining corpuses.
71
 
72
  The Dolly model family is under active development, and so any list of shortcomings is unlikely to be exhaustive, but we include known limitations and misfires here as a means to document and share our preliminary findings with the community.
73
- In particular, `dolly-v2-6-9b` struggles with: syntactically complex prompts, programming problems, mathematical operations, factual errors,
74
  dates and times, open-ended question answering, hallucination, enumerating lists of specific length, stylistic mimicry, having a sense of humor, etc.
75
- Moreover, we find that `dolly-v2-6-9b` does not have some capabilities, such as well-formatted letter writing, present in the original model.
76
 
77
  ### Dataset Limitations
78
- Like all language models, `dolly-v2-6-9b` reflects the content and limitations of its training corpuses.
79
 
80
  - **The Pile**: GPT-J’s pre-training corpus contains content mostly collected from the public internet, and like most web-scale datasets,
81
  it contains content many users would find objectionable. As such, the model is likely to reflect these shortcomings, potentially overtly
82
  in the case it is explicitly asked to produce objectionable content, and sometimes subtly, as in the case of biased or harmful implicit
83
  associations.
84
 
85
- - **`databricks-dolly-15k`**: The training data on which `dolly-v2-6-9b` is instruction tuned represents natural language instructions generated
86
  by Databricks employees during a period spanning March and April 2023 and includes passages from Wikipedia as references passages
87
  for instruction categories like closed QA and summarization. To our knowledge it does not contain obscenity, intellectual property or
88
  personally identifying information about non-public figures, but it may contain typos and factual errors.
@@ -95,7 +95,7 @@ maximize the potential of all individuals and organizations.
95
  ### Benchmark Metrics
96
 
97
  Below you'll find various models benchmark performance on the [EleutherAI LLM Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness);
98
- model results are sorted by geometric mean to produce an intelligible ordering. As outlined above, these results demonstrate that `dolly-v2-6-9b` is not state of the art,
99
  and in fact underperforms `dolly-v1-6b` in some evaluation benchmarks. We believe this owes to the composition and size of the underlying fine tuning datasets,
100
  but a robust statement as to the sources of these variations requires further study.
101
 
@@ -105,11 +105,11 @@ but a robust statement as to the sources of these variations requires further st
105
  +-----------------------------------+--------------+------------+--------------+-------------+-----------------+----------+----------+----------|
106
  | EleutherAI/pythia-2.8b | 0.348 | 0.585859 | 0.589582 | 0.591217 | 0.323379 | 0.73395 | 0.638226 | 0.523431 |
107
  | EleutherAI/pythia-6.9b | 0.368 | 0.604798 | 0.608524 | 0.631548 | 0.343857 | 0.761153 | 0.6263 | 0.543567 |
108
- | databricks/dolly-v2-2-8b | 0.384 | 0.611532 | 0.589582 | 0.650767 | 0.370307 | 0.742655 | 0.575535 | 0.544886 |
109
  | EleutherAI/pythia-12b | 0.364 | 0.627104 | 0.636148 | 0.668094 | 0.346416 | 0.760065 | 0.673394 | 0.559676 |
110
  | EleutherAI/gpt-j-6B | 0.382 | 0.621633 | 0.651144 | 0.662617 | 0.363481 | 0.761153 | 0.655963 | 0.565936 |
111
  | databricks/dolly-v2-12b | 0.408 | 0.63931 | 0.616417 | 0.707927 | 0.388225 | 0.757889 | 0.568196 | 0.56781 |
112
- | databricks/dolly-v2-6-9b | 0.392 | 0.633838 | 0.607735 | 0.686517 | 0.406997 | 0.750816 | 0.644037 | 0.573487 |
113
  | databricks/dolly-v1-6b | 0.41 | 0.62963 | 0.643252 | 0.676758 | 0.384812 | 0.773667 | 0.687768 | 0.583431 |
114
  | EleutherAI/gpt-neox-20b | 0.402 | 0.683923 | 0.656669 | 0.7142 | 0.408703 | 0.784004 | 0.695413 | 0.602236 |
115
  +-----------------------------------+--------------+------------+--------------+-------------+-----------------+----------+----------+----------+
 
5
  library_name: transformers
6
  inference: false
7
  ---
8
+ # dolly-v2-7b Model Card
9
  ## Summary
10
 
11
+ Databricks’ `dolly-v2-7b`, an instruction-following large language model trained on the Databricks machine learning platform
12
  that is licensed for commercial use. Based on `pythia-6.9b`, Dolly is trained on ~15k instruction/response fine tuning records
13
  [`databricks-dolly-15k`](https://github.com/databrickslabs/dolly/tree/master/data) generated
14
  by Databricks employees in capability domains from the InstructGPT paper, including brainstorming, classification, closed QA, generation,
15
+ information extraction, open QA and summarization. `dolly-v2-7b` is not a state-of-the-art model, but does exhibit surprisingly
16
  high quality instruction following behavior not characteristic of the foundation model on which it is based.
17
 
18
  **Owner**: Databricks, Inc.
19
 
20
  ## Model Overview
21
+ `dolly-v2-7b` is a 6.9 billion parameter causal language model created by [Databricks](https://databricks.com/) that is derived from
22
  [EleutherAI’s](https://www.eleuther.ai/) [Pythia-6.9b](https://huggingface.co/EleutherAI/pythia-6.9b) and fine-tuned
23
  on a [~15K record instruction corpus](https://github.com/databrickslabs/dolly/tree/master/data) generated by Databricks employees and released under a permissive license (CC-BY-SA)
24
 
 
32
  ```
33
 
34
  The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline`
35
+ found in the model repo [here](https://huggingface.co/databricks/dolly-v2-7b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required.
36
  Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality.
37
  It is also fine to remove it if there is sufficient memory.
38
 
 
40
  import torch
41
  from transformers import pipeline
42
 
43
+ generate_text = pipeline(model="databricks/dolly-v2-7b", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto")
44
  ```
45
 
46
  You can then use the pipeline to answer instructions:
 
49
  generate_text("Explain to me the difference between nuclear fission and fusion.")
50
  ```
51
 
52
+ Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-7b/blob/main/instruct_pipeline.py),
53
  store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer:
54
 
55
  ```
56
  from instruct_pipeline import InstructionTextGenerationPipeline
57
  from transformers import AutoModelForCausalLM, AutoTokenizer
58
 
59
+ tokenizer = AutoTokenizer.from_pretrained("databricks/dolly-v2-7b", padding_side="left")
60
+ model = AutoModelForCausalLM.from_pretrained("databricks/dolly-v2-7b", device_map="auto")
61
 
62
  generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer)
63
  ```
 
66
  ## Known Limitations
67
 
68
  ### Performance Limitations
69
+ **`dolly-v2-7b` is not a state-of-the-art generative language model** and, though quantitative benchmarking is ongoing, is not designed to perform
70
  competitively with more modern model architectures or models subject to larger pretraining corpuses.
71
 
72
  The Dolly model family is under active development, and so any list of shortcomings is unlikely to be exhaustive, but we include known limitations and misfires here as a means to document and share our preliminary findings with the community.
73
+ In particular, `dolly-v2-7b` struggles with: syntactically complex prompts, programming problems, mathematical operations, factual errors,
74
  dates and times, open-ended question answering, hallucination, enumerating lists of specific length, stylistic mimicry, having a sense of humor, etc.
75
+ Moreover, we find that `dolly-v2-7b` does not have some capabilities, such as well-formatted letter writing, present in the original model.
76
 
77
  ### Dataset Limitations
78
+ Like all language models, `dolly-v2-7b` reflects the content and limitations of its training corpuses.
79
 
80
  - **The Pile**: GPT-J’s pre-training corpus contains content mostly collected from the public internet, and like most web-scale datasets,
81
  it contains content many users would find objectionable. As such, the model is likely to reflect these shortcomings, potentially overtly
82
  in the case it is explicitly asked to produce objectionable content, and sometimes subtly, as in the case of biased or harmful implicit
83
  associations.
84
 
85
+ - **`databricks-dolly-15k`**: The training data on which `dolly-v2-7b` is instruction tuned represents natural language instructions generated
86
  by Databricks employees during a period spanning March and April 2023 and includes passages from Wikipedia as references passages
87
  for instruction categories like closed QA and summarization. To our knowledge it does not contain obscenity, intellectual property or
88
  personally identifying information about non-public figures, but it may contain typos and factual errors.
 
95
  ### Benchmark Metrics
96
 
97
  Below you'll find various models benchmark performance on the [EleutherAI LLM Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness);
98
+ model results are sorted by geometric mean to produce an intelligible ordering. As outlined above, these results demonstrate that `dolly-v2-7b` is not state of the art,
99
  and in fact underperforms `dolly-v1-6b` in some evaluation benchmarks. We believe this owes to the composition and size of the underlying fine tuning datasets,
100
  but a robust statement as to the sources of these variations requires further study.
101
 
 
105
  +-----------------------------------+--------------+------------+--------------+-------------+-----------------+----------+----------+----------|
106
  | EleutherAI/pythia-2.8b | 0.348 | 0.585859 | 0.589582 | 0.591217 | 0.323379 | 0.73395 | 0.638226 | 0.523431 |
107
  | EleutherAI/pythia-6.9b | 0.368 | 0.604798 | 0.608524 | 0.631548 | 0.343857 | 0.761153 | 0.6263 | 0.543567 |
108
+ | databricks/dolly-v2-3b | 0.384 | 0.611532 | 0.589582 | 0.650767 | 0.370307 | 0.742655 | 0.575535 | 0.544886 |
109
  | EleutherAI/pythia-12b | 0.364 | 0.627104 | 0.636148 | 0.668094 | 0.346416 | 0.760065 | 0.673394 | 0.559676 |
110
  | EleutherAI/gpt-j-6B | 0.382 | 0.621633 | 0.651144 | 0.662617 | 0.363481 | 0.761153 | 0.655963 | 0.565936 |
111
  | databricks/dolly-v2-12b | 0.408 | 0.63931 | 0.616417 | 0.707927 | 0.388225 | 0.757889 | 0.568196 | 0.56781 |
112
+ | databricks/dolly-v2-7b | 0.392 | 0.633838 | 0.607735 | 0.686517 | 0.406997 | 0.750816 | 0.644037 | 0.573487 |
113
  | databricks/dolly-v1-6b | 0.41 | 0.62963 | 0.643252 | 0.676758 | 0.384812 | 0.773667 | 0.687768 | 0.583431 |
114
  | EleutherAI/gpt-neox-20b | 0.402 | 0.683923 | 0.656669 | 0.7142 | 0.408703 | 0.784004 | 0.695413 | 0.602236 |
115
  +-----------------------------------+--------------+------------+--------------+-------------+-----------------+----------+----------+----------+